• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

6B4G SRPP

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
G

GK

Disabled Account
Joined 2006
6b4g Srpp

I’ve got a box of low power (5Wrms) PA speaker transformers with tapped primary impedances ranging 500-5000 ohms and secondary taps for 4,8 and 16 ohms which I bought a while ago for about 5 bucks a pop. I’ve tested them out and they work rather well – they are surprisingly flat out to 20kHz and quite happily reproduce asymmetrical audio signals with little distortion.
Problem is they can’t take the specified 60mA quiescent plate current of a single-ended 6B4G without saturating and they are not suitable for push-pull.
I’ve got a dozen or so 6B4G’s, so I’m thinking of making a stereo amplifier using a pair of them for each channel, SRPP connected with a regulated 500V plate supply (I have a few dozen 12E1 series pass tetrodes, so a few of these will come in handy here). The output will be capacitivly coupled to the transformer primary, thus circumventing the DC saturation problem.

So who do I calculate the optimal (Pout max) load resistance for a pair of 6B4G’s in SRPP then?

PS
And yes, I do know that this amplifier will be ridiculously inefficient with a rather low power output, but I don’t care.


Cheers,
Glen
 
G

GK

Disabled Account
Joined 2006
Thanks for the link. I’ve had a quick peruse of a few SRPP-topic blogs on that site so far, but they don’t address what I’m trying to do. I’ll keep searching.

Anyway, attached below is a basic diagram of what I plan to do. An individual floating regulated DC supply powers the heater of each tube.


Cheers,
Glen
 

Attachments

  • srpp.gif
    srpp.gif
    43.6 KB · Views: 809
G

GK

Disabled Account
Joined 2006
Tweeker said:
A 6B4G SRPP should be able to manage 5+ watts out.

The totem-pole Amplifier. Taub&Millman, Pulse and Digital Circuit, Cap.3, Linear Pulse Amplifiers, pp. 99-101, McGraw-Hill, 1956. This has the math you need.
Click to expand...


Tweeker said:
Is a 590V supply possible, you have cathode bias to account for. Theres also a fixed biased variant you could rather use.
Click to expand...


Tweeker said:
For a 500V supply and cathode bias I come up with this:
Vbias=34V, Ik=67ma, Rk=510ohms, mu=4.2, gm=5350mhos.
Anode=216V, plate dissipation=14.5watts.

Rk=1/gm+2Rl/mu

Optimum load =~680ohms.
Click to expand...


Fantastic! Thanks for that. You're right about the plate voltage. I also made an error on my scribble - the anode of the lower tube idles at +250V, not the cathode of the top tube. With cathode resistor biasing and a +500V supply, both tubes effectively run with a plate voltage of 205V. This should still work fine, but in order to get the most out of the tubes, I'll run +545V with fixed bias on the bottom tube. That will give me 250V across each tube and will do away with the need for a large bypass capacitor.

Now I can log-off and start calculating the load impedance :)

Cheers,
Glen
 
T

Tweeker

Member
Joined 2004
For 60ma bias resistor is actually around 583ohms, and anode is at about 215V. 35V is across the bias resistor, which is lost from the anode voltage.

The fixed bias variant requires some other changes, linked article details. I think a mixed bias version might not push-pull as evenly.
 
M

martinab2

Member
Joined 2004
Your 12e1 tubes would also make a reasonable SRPP output stage. Being series pass tubes they are happier at lower voltage and higher current than the 6b4s. I built one which used 450v ht and 120ma (still well within the limits of the 12e1).

If you are looking at SRPP, you should also consider the Aikido which makes a good output stage as well as a preamp. I prefer the sound over the SRPP although it does have a bit less gain.
 
Tubes4e4

Tubes4e4

Member
Joined 2005
Hi Manta,

Frequency response is amazing !
Click to expand...

Yes, but I also gave a lot of warnings and considerations. If one is not willing and able to invest considerable efforts in buying toroids and measuring through them, this combination is a game of chance, IMNSHO. Serge with his KT88SRPP had pure luck that he succeeded so well with the first toroid he tried.

Everyone reading this has been warned.

Regards,

Tom
 
G

GK

Disabled Account
Joined 2006
Tweeker said:
For 60ma bias resistor is actually around 583ohms, and anode is at about 215V. 35V is across the bias resistor, which is lost from the anode voltage.

The fixed bias variant requires some other changes, linked article details. I think a mixed bias version might not push-pull as evenly.
Click to expand...



Tweeker said:
For 60ma bias resistor is actually around 583ohms, and anode is at about 215V. 35V is across the bias resistor, which is lost from the anode voltage.

The fixed bias variant requires some other changes, linked article details. I think a mixed bias version might not push-pull as evenly.
Click to expand...


Okay.......I've been hitting the theory and running a few sims and I reckon I've got this circuit sussed out :D :xeye:

RCA datasheet for the octal 6B4G (and its 4 pin equivalent, the 6A3) give the following values for single ended class A, fixed bias operation:

Vplate = 250V
Vg = -45V
Ia = 60mA
umhos = 5250
u=4.2
Ri = 800 ohms
Rload = 2500 ohms
Pout = 3.2W


For cathode resistor bias, this equates to a cathode resistor of 45/0.06 = 750 ohms, so long as the plate voltage is increased by 45V to 295V to compensate for the biasing voltage lost across the cathode resistor.

Fixed biasing for the lower tube doesn't have any effect on the push-pull operation of the SRPP circuit. This is governed entirely by voltage developed across the cathode resistor of the upper tube.
So long as the load current is taken from the cathode of the top tube, the circuit is inherently balanced.
The textbook “totem pole amplifier” described in the first link you gave me is in fact based on fixed bias for the lower tube.

I will run the lower tube with a fixed bias voltage of –45V with the cathode grounded. It will have an anode current of 60mA and the plate will idle at 250V. The top tube will have a 750 ohm cathode resistor to give an identical grid bias of –45V. This means that the cathode will sit at 295V. The plate supply for the upper tube will therefore be 545V, so that both tubes run balanced with 250V anode-cathode.

With regards to my opening question, namely the optimal load resistance, the answer should have been obvious – it is exactly half the recommended value for a single tube. For a single 6B4G the specified load resistance is 2500 ohms for a power output of 3.2W. For two 6B4G’s in SRPP, it is simply 1250 ohms for a power output of 6.4W.

In the SRPP circuit, the peak load current is effectively doubled over that deliverable by a single tube. This is because there is now a tube on the top operating out of phase. For instance, when the lower tube is driven to conduct and extra 10mA from the quiescent state, the upper tube is driven to conduct 10mA less from the quiescent state. The difference of 20mA therefore flows through the load. This is why the load resistance needs to be halved – so that the voltage swings on each tube exactly mimics that of a single tube amplifier.

The output resistance for a triode SRPP with fixed bias for the bottom tube (or bypassed cathode resistor bias) is equal to:

Ri(Ri+Rk)/(2Ri+Rk(u+1))

Ri = the tubes internal resistance (800 ohms for the 6B4G)
Rk = the upper tubes cathode resistor (750 ohms)
U = amplification factor (4.2)

For my 6B4G circuit, this = 800(800+750)/((1600+(750*5.2)) = 225ohms
Since the load resistance is 1250 ohms, the damping factor is 1250/225 = 5.6. This is reasonable improvement over, say two 6B4G’s in parallel to give 6.4W into 1250 ohms, whose damping factor would be Rload/(Ri/2) = 1250/(800/2) = 3.1.

Cheers,
Glen
 
G

GK

Disabled Account
Joined 2006
martinab2 said:
Schematic for my old 12e1 SRPP amp (now rewired as Aikido with 6BL7s for first stage)
An externally hosted image should be here but it was not working when we last tested it.
Click to expand...


Thanks for the idea. I'm going to need a large proportion of my NOS 12E1 stock to get a couple of Marconi regulated laboratory power supplies up and runing, so I'll stick with the 6B4G's for now though.
Maybe a future project :)

Cheers,
Glen
 
Tubes4e4

Tubes4e4

Member
Joined 2005
Hi Glen,

With regards to my opening question, namely the optimal load resistance, the answer should have been obvious – it is exactly half the recommended value for a single tube.
Click to expand...

Here are some (empirical) findings you might be interested in. Note shifting of optimum reflected load value when gNFB is applied :att'n:

ECL82/PCL82/6BM8/16A8 triode SRPP load matching table (plate resistance for triode strapped pentode section is about 1600 ohms):

An externally hosted image should be here but it was not working when we last tested it.


ECL85/PCL85/6GV8/18GV8 triode SRPP load matching table (plate resistance for triode strapped pentode section is about 1400 ohms):

An externally hosted image should be here but it was not working when we last tested it.



Hope it helps,

Tom

EDIT: Tag line adjusted accordingly ;)
 
G

GK

Disabled Account
Joined 2006
Tubes4e4 said:
Hi Glen,



Here are some (empirical) findings you might be interested in. Note shifting of optimum reflected load value when gNFB is applied :att'n:

ECL82/PCL82/6BM8/16A8 triode SRPP load matching table (plate resistance for triode strapped pentode section is about 1600 ohms):

An externally hosted image should be here but it was not working when we last tested it.


ECL85/PCL85/6GV8/18GV8 triode SRPP load matching table (plate resistance for triode strapped pentode section is about 1400 ohms):

An externally hosted image should be here but it was not working when we last tested it.



Hope it helps,

Tom

EDIT: Tag line adjusted accordingly ;)
Click to expand...


G’day Tom.

The optimal load resistance for the triode SRPP has a relationship to the cathode biasing resistor for the upper tube as dictated by the formula:

Rcathode=1/gm+(2Rload/u), as described by Taub&Millman here:

http://digilander.libero.it/paeng/the_totem_pole_amplifier_paragraph.htm

The datasheet for the 6B4G gives an optimal biasing point for class A operation (in terms of power output and distortion) of 60mA with a plate voltage of 250V, which requires a grid voltage of approximately –45V. The recommended load resistance is 2500 ohms.
Running two 6B4G triodes in SRPP, each with the recommended 60mA plate current and 250V plate voltage dictates the value of the upper tube’s cathode biasing resistor – In this case, approximately 750 ohms.
In SRPP, each tube only effectively sees a load resistance of half the actual, since each tube is sharing the current drive to the load with the other. Therefore, in order to reflect the recommended load impedance to each tube, the physical load value needs to be halved to 1250 ohms, so that each tube ‘sees’ the recommended 2500 ohms.

This can be proven by applying the above formula, plugging in 1250 ohms for Rload:

Rcathode=1/gm+(2Rload/u)

=Rcathode=1/0.00525+(2*1250/4.2) = 785 ohms = ~750 ohms.

Also, I’m not sure what you mean by “Note shifting of optimum reflected load value when gNFB is applied ”

How was the gNFB applied for those calculations? Applying global negative feedback, say from the speaker output to the cathode of a valve driving the SRPP output stage will lower the amplifiers output impedance and improve the damping factor, but it cannot alter the load impedance reflected back to the SRPP output stage.

Cheers,
Glen
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom