Bias voltage by Cap

[XTRXTR]

Carvin Legacy 100W Model has an odd bias circuit.
I was watching Stuart the UK guitar amp guy and wondered if there are other amps that use this type of circuit?

I haven't seen an AC bias voltage reduced by a series cap before.
https://www.thetubestore.com/lib/th...arvin/Carvin-Legacy-Vl100-Vl212-Schematic.pdf

 

[mickeydg5]

It is a fairly typical negative bias supply with the addition of a switched bias level circuit for both 6L6/KT66 and 6CA7/EL34/KT77 types.

 

[sunflower]

JCM900 has same voltage dropping capacitor in bias circuit.
https://www.thetubestore.com/lib/th...arshall-JCM900-Dualrev-50W-4500-Schematic.pdf

It seems like not a good design.
https://www.marshallforum.com/threads/jcm900-4500-el34-red-plating-run-away-bias-current.122500/

 

[351]

the design is good, cheap caps are not.

 

[sunflower]

I found an article. I didn't know about " capacitor coupled bias supply ".
http://www.valvewizard.co.uk/bias.html

 

[XTRXTR]

JCM900 has same voltage dropping capacitor in bias circuit.
https://www.thetubestore.com/lib/th...arshall-JCM900-Dualrev-50W-4500-Schematic.pdf

It seems like not a good design.
https://www.marshallforum.com/threads/jcm900-4500-el34-red-plating-run-away-bias-current.122500/

Thanks for those links.
I see the same bias circuit is used and the thread OP suggests the same type of run away bias on his JCM900 that the owner is experiencing with the Carvin Legacy 100W in the video I linked. Even Merlin Blencowe states that bias circuit should not be used for anything greater than a 50W due to issues with the grid leak and bias R2 total resistance max stated on the data sheets. I wonder if that can be tweaked for larger Watt amps by adjusting the R1 and R2 values as well as the Leak resistors for the particular amp/tubes design? Further the cap should be rated for more than the full HT voltage or higher and it should be good quality plastic/poly type. Looks like the JCM900 suffered a failure (short) of the coupling cap and likely not rated high enough. The wizard also hints at the need for a 2w bleed resistor on the reservoir cap for this type of circuit. I don't see that on the JCM900 schematic provided but the schem is perhaps not at the revision level it is as built.
Great information sunflower

 

[Pete Farrington]

Even Merlin Blencowe states that bias circuit should not be used for anything greater than a 50W due to issues with the grid leak and bias R2 total resistance max stated on the data sheets

The bigger you make R2 (the bias adjust circuit), the greater the available negative voltage. It is a good idea to make it larger than R1 otherwise loading becomes excessive. However, it can’t be too large since it forms part of the output valves’ grid leak path. This also makes this sort of bias supply unsuitable for powerful amplifiers with big/multiple output valves that need a low-impedance grid leak path.

Most valve guitar amps don’t bother with keeping within the limiting value of grid circuit resistance.

 

[mickeydg5]

@XTRXTR

To answer the title question, the series capacitor works in conjunction with the first resistor to ground achieving an adequate voltage level (drop) for the bias supply. Actually the entire bias supply circuit works together to achieve this but has adjustability built in for convenience.

The series capacitor does have internal resistance. It is called Xc. For this capacitor operating at about 60Hz (of the power supply) the resistance will be in the 56k range.

Tell Merlin not to worry to much because Carvin uses less than 245k of the rated 700k for control grid (Grid No.1) circuit resistance.

 

[66 Kicks]

JCM900 has same voltage dropping capacitor in bias circuit.

The JCM900 bias circuit, aka Ampeg bias circuit, is a totally different animal. The Carvin bias circuit uses the impedance of the .047uF capacitor in series with the resistance of the 100K resistor to form a voltage divider. You can get nearly the same result by replacing the .047uF capacitor with a 68K resistor for a purely resistive voltage divider. Or you could do away with the voltage divider altogether and just put an appropriate resistor between the HT and the diode like everybody else does. The JCM900 bias circuit must have the .047uF capacitor in order to function.

 

[66 Kicks]

@XTRXTR

To answer the title question, the series capacitor works in conjunction with the first resistor to ground achieving an adequate voltage level (drop) for the bias supply. Actually the entire bias supply circuit works together to achieve this but has adjustability built in for convenience.

The series capacitor does have internal resistance. It is called Xc. For this capacitor operating at about 60Hz (of the power supply) the resistance will be in the 56k range.

Tell Merlin not to worry to much because Carvin uses less than 245k of the rated 700k for control grid (Grid No.1) circuit resistance.

Unfortunate choice of terminology and symbol, but I know what you meant.

The worst case resistance that I see in the bias circuit is 42K. The current from four tubes flows through this resistance, so the Effective resistance is 4 x 42K=168K. Two tubes are connected to each grid leak, so the Effective resistance is 2 x 220K=440K. The total Effective grid leak resistance is 168K + 440K = 608K. If anybody agrees with this, they might want to hold off on telling Merlin anything.

 

[Pete Farrington]

Carvin uses less than 245k of the rated 700k for control grid (Grid No.1) circuit resistance

Bear in mind that the 220k grid leaks are shared between 2 grids. So perhaps equivalent to 440k?
Then another 40k ish shared between 4 grids, so that section adds a equivalent of 160k per grid.

So a total of 600k per grid.

 

[Pete Farrington]

Carvin uses less than 245k of the rated 700k for control grid (Grid No.1) circuit resistance

Bear in mind that the 220k grid leaks are shared between 2 grids. So perhaps equivalent to 440k?
Then another 40k ish shared between 4 grids, so that section adds a equivalent of 160k per grid.

So a total of 600k per grid.

 

[alpha al]

Ampegs like the V2, V4 and Gemini series used a cap coupled bias supply. I never liked that design. They did it to save money by not needing a power transformer with a separate bias winding.

 

[66 Kicks]

Ampegs like the V2, V4 and Gemini series used a cap coupled bias supply. I never liked that design. They did it to save money by not needing a power transformer with a separate bias winding.

That's where Marshall stole the circuit from. One of those dates back to 1967.

I never much liked that design, either. I prefer the Vox one. You know, the one that shunts the positive side of the AC to ground via a diode.

 

[sunflower]

The JCM900 bias circuit, aka Ampeg bias circuit, is a totally different animal. The Carvin bias circuit uses the impedance of the .047uF capacitor in series with the resistance of the 100K resistor to form a voltage divider. You can get nearly the same result by replacing the .047uF capacitor with a 68K resistor for a purely resistive voltage divider. Or you could do away with the voltage divider altogether and just put an appropriate resistor between the HT and the diode like everybody else does. The JCM900 bias circuit must have the .047uF capacitor in order to function.

Hi,
Thank you for correcting and providing further information.

 

[66 Kicks]

Hi,
Thank you for correcting and providing further information.

Here is some more information:

 

[66 Kicks]

And some more:

 

[mickeydg5]

Unfortunate choice of terminology and symbol, but I know what you meant.

The worst case resistance that I see in the bias circuit is 42K. The current from four tubes flows through this resistance, so the Effective resistance is 4 x 42K=168K. Two tubes are connected to each grid leak, so the Effective resistance is 2 x 220K=440K. The total Effective grid leak resistance is 168K + 440K = 608K. If anybody agrees with this, they might want to hold off on telling Merlin anything.

Bear in mind that the 220k grid leaks are shared between 2 grids. So perhaps equivalent to 440k?
Then another 40k ish shared between 4 grids, so that section adds a equivalent of 160k per grid.

So a total of 600k per grid.

This aint no cathode bias circuit impeding the tube operation/overall current.
You guys seem to think what is wanted is grid current. Its the opposite, less grid current.
What is being talked about here is the bias network as associated with the grid-cathode circuit.
The purpose for the maximum control grid rating is for proper coupling to the control grid. The purpose is to curtail the input strength/power.
The load resistor to ground in the grid-cathode circuit is not just part of the control grid but also part of the preceding stage controlling its gain.

When two tubes are placed in parallel the signal power is divided in half between the two tubes.
Voltage to each is the same. That voltage is the driving factor, both DC and AC.

The so called "leak" or load resistor is leaking signal to ground in order to curtail signal and control the grid itself and therefore the tube operations.

Too much positive grid voltage is bad at a point but more so too much negative grid voltage is bad period. Negative current, that from grid to cathode, is to be avoided.

Everyone, tell me if these control grid resistance values matter so much why are there not more differences between typical same tube type 50W (2 tubes) and 100W (4 tubes) and so on? Think about it, they are practically the same.

By the way, I like my words and symbols.

 

[66 Kicks]

This aint no cathode bias circuit impeding the tube operation/overall current.
You guys seem to think what is wanted is grid current. Its the opposite, less grid current.
What is being talked about here is the bias network as associated with the grid-cathode circuit.
The purpose for the maximum control grid rating is for proper coupling to the control grid. The purpose is to curtail the input strength/power.
The load resistor to ground in the grid-cathode circuit is not just part of the control grid but also part of the preceding stage controlling its gain.

When two tubes are placed in parallel the signal power is divided in half between the two tubes.
Voltage to each is the same. That voltage is the driving factor, both DC and AC.

The so called "leak" or load resistor is leaking signal to ground in order to curtail signal and control the grid itself and therefore the tube operations.

Too much positive grid voltage is bad at a point but more so too much negative grid voltage is bad period. Negative current, that from grid to cathode, is to be avoided.

Everyone, tell me if these control grid resistance values matter so much why are there not more differences between typical same tube type 50W (2 tubes) and 100W (4 tubes) and so on? Think about it, they are practically the same.

By the way, I like my words and symbols.

Boy, I'm glad now that pdf64 repeated what I said, otherwise I would have been taking this tongue-lashing all by myself.

I overlooked the 4.7K grid stoppers. It's up to 612.7K now.

 

[Pete Farrington]

You guys seem to think what is wanted is grid current

Sorry for the misunderstanding, but I don’t see where anyone has put that viewpoint forward?

The purpose for the maximum control grid rating is for proper coupling to the control grid

The grid leak resistance acts to tie the control grid to the operating point.
The larger the resistance in the grid circuit, the looser the tie, and the more susceptible the operating point will be to residual grid current. The value of the grid leak resistance constrains the grid voltage that can result from grid current.
Grid current can be positive or negative, ie the grid can capture electrons or emit them.
The problem is positive grid current. This may occur when valve dissipation is very high and the grid becomes hot enough for it to exhibit thermionic emission. That acts to move the operating point hotter, which can set off a vicious cycle of increasing grid emission increasing dissipation increasing grid emission … leading to valve meltdown.
Hence the grid resistance limit typically being lower in fixed bias than self (cathode) bias, as the latter will tend to counteract the above cycle.
At high output levels, all output valves will be dissipating heavily simultaneously, and hence all may exhibit high positive grid current simultaneously. So the grid resistance per valve needs factoring in, can’t just ignore it.
And whilst 700k grid resistance may be fine for a set of good, fresh EL34, older valves can have increased grid current (dunno why?), and can be prone to meltdown, especially in 100W amps.

f these control grid resistance values matter so much why are there not more differences between typical same tube type 50W (2 tubes) and 100W (4 tubes) and so on?

Because valve guitar amps tend to be ‘designed by Bubba’?
And valves are fairly resilient.
But also if valve dissipation can be limited, eg by reasonable bias setting, then the grid circuit resistance limit can be relaxed, eg see MO KT66 info https://tubedata.altanatubes.com.br/sheets/086/k/KT66.pdf