I want to use elevated heaters in a Marshall amp. Help.

[cluster]

Hello All,

I want to use elevated heaters in a Marshall High Gain Amp Design. How do I do it? Would I attached the CT of the filament to an un-used Positive Lug of a Can Cap with a resistor going to ground to lower the voltage. Would I need a balancing resistor between the 2 pos lugs if it were a dual can like a 16/16uF 450VDC? What formula would I use to determine the value of the resistors?? Can anyone explain the manner in which to do this correctly? Thx

[Helmholtz]

http://www.valvewizard.co.uk/heater.html

[strato56]

The Marshall DSL401 uses elevated heaters.

[cluster]

Would something like this work? Green CT from Filaments

[Helmholtz]

Do not connect the heater CT directly to a high voltage supply. You need a voltage divider and an extra filter cap as shown in the Valvewizard article or used in the Marshall JVM410.

Post your schematic and layout.

[cluster]

Do not connect the heater CT directly to a high voltage supply. You need a voltage divider and an extra filter cap as shown in the Valvewizard article or used in the Marshall JVM410.

Post your schematic and layout.

Could you tell me how to calculate the resistor values to use for the divider? Ohm's law?

[cluster]

So based to a voltage divider calculator a power source of 440VDC with R1= 1M and R2=100k should give me a 40VDC elevation.

[trobbins]

Perhaps 2x 470k in series if you don't have a 1M resistor that can handle at least 400Vdc.

[Chuck H]

I'm with Helmholtz. Let's have a schematic please. I didn't peruse the Valve Wizard link (though I'm sure it's informative) but it's not hard to do. So since you have questions I think it would be best if you post a schematic and take instruction. Since this seems to me a new build I'd be inclined to just integrate the divider for the filament elevation in the HV bleed circuit.

[Gregg]

Could you tell me how to calculate the resistor values to use for the divider? Ohm's law?

https://ohmslawcalculator.com/voltag...der-calculator

Let's have a schematic please.

Schematic (JVM410) was posted above. It doesn't get more simple than that.
Sometimes simple things here are getting unnecessarily complicated and drowned theory.
In this case take the HV from screens node. Calculate elevation volts (top resistor 470k-1М)). Connect heaters' CT or artificial CT to voltage divider's node. Done.

[catalin gramada]

https://ohmslawcalculator.com/voltag...der-calculator



Schematic (JVM410) was posted above. It doesn't get more simple than that.
Sometimes simple things here are getting unnecessarily complicated and drowned theory.
In this case take the HV from screens node. Calculate elevation volts (top resistor 470k-1М)). Connect heaters' CT or artificial CT to voltage divider's node. Done.

Don't forget the filter cap...
Some basic teory still necessary to understand what you're doing.
First, choose a supply node from at least two point of view : inevitably the divider will draw some current to ground (the divider is a power consumer by itself), and we don't want to affect the voltages over the chain supply so is preferable to install as close to the power supply and not in the end of the supply chain. Second: we want a quiet supply node for our DC reference, as much free of ripple. From this point the screen supply node is good compromise. Calculate the voltage you need . Choose the values of resistors in respect with voltage you need but also how much current you want to spend. But keep in mind a large value of series resistor, like 470k-1M will drastically limit the charging current of the filter capacitor, meant you will get the nominal voltage in seconds or tens of seconds. From the same reason the filter cap should not be larger than 10-20 uF large enough to smooth the ripple for our new filament reference point. Is all compromise.(for instance: 1M resistor with 100uF cap have a time constant 100 sec.)-you have a parallel resistor here meant bigger series resistor=longer time to stabilise.
From the screen node put a 220k series with 22k shunt. Instal a 10-22uF/ 100v cap over shunt resistor and tie the filament center tap to this . Done.
Eg: Suppose you have 380v at screens 220k series/22k shunt=34.5 V , 220+22= 244k, 380/244000=1.56mA , 380-34.5=345.5V over series resistor. 345Vx0.00156A= 0.538W power dissipation. 1W resistor will be enough, but 2w better choice.
From a 440V supply node with 220k/22k you get 40V DC elevation. Just change the values in the picture you posted: replace 470k with 220k/2W and 820 ohm with 22K.
470k/820 do almost nothing. Is a minimal elevation maybe to suppress some noise due h-k coupling but is not a elevation in a sense we used to protect heater to cathodes insulation at high potential like in cathode folowers or cathodyne inverter circuits. You forgot to tell us what is the reason you intend to elevate the heaters reference? We usually elevate the heater reference for a reason. And as the side aspect- you may find slight difference in sound.
But could be a mistake as well, as we see in jvm 410 schematic posted above. It use 470k/82k and not 470k/820 ohm. They made it in a rush and was confused? You have to take time for everything when ask two grands for an amp...I think. 820 ohm is not the same as 82K of course . But maybe some of those mistakes who made a difference from a boring to a brilliant amp? It is as simple as hell...

[Gregg]

Don't forget the filter cap.

I didn't. 10uF is more than enough.

I'm with you about the divider values. Sometimes if you don't know how to do it look how the "big ones" did it and just copy it although they also make mistakes.

It use 470k/82k and not 470k/820 ohm.

It's 470k/82k (R81/R83) in the JVM410 schematic.

[tedmich]

marshall_jvm410.pdf Schematic (I couldn't find in thread) you have to click "insert" after uploading apparently : (

[Chuck H]

The OP never indicated that the jvm410 was even similar to the amp in question. It was strato56 in post #3 that posted the schematic as an example of how Marshall did it with that model. Although there may be a good solution to the OP's problem by perusing the schematic, there may not. We need more information about the actual amp in question before a solution can be arrived at.

[Enzo]

Probably late in the game, but I never liked dropping from hundreds of volts to get a few. Were it mine, I'd be inclined to add a second diode facing the other way at the bias tap and make a small low voltage positive supply for the purpose. Bonus: no way for some failure to put B+ in your heater circuit.