• 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.

Low Voltage Plate Follower 6DJ8 Buffer Power Supply

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am designing a tube buffer for a Buffered Non-Inverted Gainclone monoblock.

My design is a plate follower circuit with a gain of just over 1. I have chosen to use a B+ voltage of 70 volts.

This is my first amplifier project, and I am unfamiliar with what type of power supply would be necessary for the buffer. I have been having a hard time finding power supply schematics for such a circuit. Any help would be greatly appreciated.

My initial circuit is attached. Any comments on the circuit design would be great, as this must be completed for a college project. Thanks again for any help.
 

Attachments

  • amp schematic.gif
    amp schematic.gif
    5.1 KB · Views: 1,968
Do you really want to run the tube with that low of a plate load? Yes, the feedback will help linearize things, but you're starting with something pretty nonlinear. What's wrong with a cathode follower?

In any case, a simple supply is all you need. I generally use something like a Maida regulator as outlined in the infamous National Semiconductor LB47. You probably don't need a Darlington.
 
It looks like you are using the 1K5 resistor to introduce lots of feedback.
May I suggest that a better way to do it would be something like the JLTi gainclone, though there are better buffers than the simple CF which Joe Rasmussen suggests, unfortunately the full JTLi is a copyrighted design and its up to yourself to work out the component values.

http://www.customanalogue.com/diytubegainclone/index.htm

Shoog
 
The idea of my project was to use low voltages for the 6DJ8 tubes. It seems as if I will have to use a much higher B+ voltage for a cathode follower, and I wanted to try and have a gain just above unity. From other forums I have heard that the 6DJ8 runs well from lower supply voltages. Any suggestions for a higher B+ voltage, or even a better circuit topology for a slightly above unity gain buffer?

Also, I had a question about the wiring for the heater. I am only using one triode in the 6DJ8, and I wanted to know how to wire the heating up. The datasheets say that a parallel supply of 6.3V @ 350 mA is needed. But where do I set the reference for this voltage? to ground or the cathode? Does this mean I need 700 mA for the heaters?

Thanks again for all the help.
 
You don't need a higher voltage for a CF. The same 70V rail will work fine. Use a constant current sink in the cathode to run the tube at 6-8mA, and pull the CCS to the -35V rail you have for the 1875.

You'll need to power the entire heater of the tube, 6.3V terminal to terminal at 300-365mA (depending on the flavor of ECC88/6DJ8/whatever) that you use. The heater supply can be referenced to ground, but for lowest noise, it's better to reference it to 20-40V above ground.
 
So you are suggesting to use a CCS that feeds into the cathode from the -35V rail of the LM3875. I am not too familiar with constant current source circuits, but this does seem like a good idea. Do you know of any helpful links for CCSs or any basic schematics that I could learn from.

Also, how would you suggest to reference the heaters at 20-40V. I am somewhat confused by this, but I have seen this suggestion around the internet. Any more insight would be great.

Thanks.
 
Attached are the results for the 70V B+ cathode follower from John Broskie's TubeCad. So would I basically replace the cathode resistor with the CCS that was suggested? I am going to do a few Spice simulations to obtain some performance characteristics of the circuit. Any suggestions about the circuit or how to implement a CCS to feed the cathode would be greatly appreciated?

Thanks again for all the help from such a great group of people.
 
In order to create your +40V reference you need to just create a resistor voltage divider between +B and ground. With a 70V +B you could use 300K 400K. with the three hundred tied to +B and the 400K tied to earth. The reference is taken from the junction and tied to one side of your heater winding. The ECC88 is insensitive to heater induced hum so no other special precautions will be needed. Infact you can probably get away with just referencing to earth.

As noted in another thread the ECC88 has extremely wide bandwidth and so will tend to oscillate if not wired up correctly. Grid stoppers are essential and should be wired straight onto the valve base. If you decide on a CF and there is no load resistor, I would advise a small resistor be wired onto the anode pin and between the anode and +B, something in the order of 10 - 100 ohms.

Its easy to make the ECC88 behave well if these precautions are observed.

Shoog
 
mwl6m said:
So you are suggesting to use a CCS that feeds into the cathode from the -35V rail of the LM3875. I am not too familiar with constant current source circuits, but this does seem like a good idea. Do you know of any helpful links for CCSs or any basic schematics that I could learn from.

Also, how would you suggest to reference the heaters at 20-40V. I am somewhat confused by this, but I have seen this suggestion around the internet. Any more insight would be great.

For free, and worth every penny, do a search on this site for my "Heretical" line amp. For money, and worth much more, grab a copy of "Valve Amplifiers" 3rd edition (Morgan Jones).
 
I have 2 25-0-25V secondary center tapped transformers that I can use for my power supply(s). Any suggestions of how to obtain an output voltage of +70V from one (or both) of these. I need the 70 volts for my buffer, but I am very new in the area of power supply design. Something simple will work best for me.

I am also designing the power supply for my gainclone amp. I am planning on using +/-35V and would like to be able to obtain this and the +70V from the two transformers. Any help will be greatly appreciated. Thanks ahead of time for the help.
 
Hi mwl6m ,

In my opinion , you must to do the following :

1 – With one transformer and an ordinary silicon bridge ,
you can get the split power supply , with ( aprox. ) + 34 V
and – 34 V using the center tap as the reference zero
( ground ) , it is a classic design . The two 1500 uf x 50 V
elec. caps , that you showed on your design are more than
enough . It is more reliable to by pass each one with a 0.1
uf x 250 V polyester cap , very close to the LM 3875 pins .
This splited PS will feed the LM 3875 .

2 – With another transformer , you have to implement a
voltage doubler , using only the two side taps ( 50 Vac )
( forget and isolate the center tap ) , with two silicon diodes
1N4007 and two elec. caps of 100 uf x 100V , and then
you will get an output voltage ( at the filter input ) around
135 V , then connect the voltage doubler to a 1K2 x 0.5 W
carbon resistor , then to a 220 uf x 200 V cap , then to a 8K2
x 1 W resistor and finally to a 47 uf x 100 V cap , where
will be the 70 / 75 Volts , very well filtered , without any trace
of “ hummm “ .
I ‘ve calculated the values above , regarding the 6DJ8 , 6 mA
cathode ( and plate ) current .
Any doubt , feel free to ask .


Regards ,

Carlos
 
I have been working with wiring up the tube's heater. I have a lab power supply that I have been using to try and test the tube, but when I apply the 6.3V, there is no current being drawn by the heaters. I have the heaters hooked up in parallel and referenced to ground, but it does not seem to want to work. Do I need some kind of resistor divider network to get the current I need, or am I just missing something obvious? Does the tube glow when the heater voltage is applied, or only after the B+ voltage is applied? Any help would be greatly appreciated. Thanks for the help.
 
I get 4.6 ohms on a Sovtek 6922 and 4.7 ohms on a EH6922.
So I guess yours must be OK. With that kind of cold resistance it seems impossible to have no heater current when you apply a voltage on those pins. You are looking at pin 4 and 5 from the bottom , reading clockwise .
Cheers.
 
Hi mwl6m ,

Are you sure that your lab power supply is able to
supply the 0.72 Ampere that two 6DJ8 paralleled heaters
need to work properly ( 0.60 Ampere for two 6922 ) ???
May be your lab power supply does have any kind
of short circuit protection , that lowers the output
voltage to aprox. ZERO , when you connect it across
the two COLD heaters . Remember that the current
demand of two heaters when COLD , are many times
higher than the nominal working current ( hot heaters ) .
A COLD heater seems like a short circuit from the
the power supply output point of view !!

Furthermore , are you sure about the ASHOK advice ??

“You are looking at pin 4 and 5 from the bottom , reading
clockwise ?? “

Double check , and feedback ,

Regards ,

Carlos
 
I had the heaters wired wrong, but I got them to work. When I tested the tube circuit with a sin wave input of amplitude 1 volt it seemed to work fine, but when it is connected to the gainclone output stage, it doesn't seem to work. any suggestions? Do i need grid leak resistors in the tube circuit. Any suggestions are appreciated. Thanks.
 
A few points to consider.
In your circuit the grid of the tube goes to the slider of the pot. Now as you turn the pot it is possible for the slider to momentarily leave the track due to dust or other particles. That will cause the grid to float and voltage at the anode to jump to the supply voltage. Not a good thing at all. So a fixed resistor from gate to ground is probably mandatory to avoid any possibility of this happening. 1 Meg should be fine.

Additionally the LM3875 input also goes to the same resistor and can cause similar problems with the power amp chip. The fluctuating voltage at the anode will also appear momentarily at the input of the 3875 and if after the resistor drops it is still too high , the chip will be 'fried'. Positive input to ground resistor is essential for the 3875. Maybe your 3875 is already dead ? Does it work by itself directly from the source ? You would have confirmed this of course.

You have shunt feedback on your tube and a feedback resistor of 1.5K and 1K is totally out of line for this tube stage ! Try 47K at the input and 68K or 75K from output to the grid. It will degrade the s/n ratio but it should still be pretty good for your circuit ( 1 v input ). The resistor's contribution would give a s/n ratio of about
just over 107 db wrt to 1 volt ( ignoring all other noise sources).

Do not depend on the feedback and grid circuit as a bias path for the input of the 3875. Use a resistor from + input to ground ( 22K or more ).

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