• 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.

GU-50 in Single-Ended Triode Configuration - What bias points?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all audio fanatic friends,

This may be a silly question, so please forgive me for asking as I am still a NEWBIE to tube amps

I have constructed an RH84 amp with the 12AT7 and 6P14P (Russian EL84 equivalent) in triode mode and find the sound really nice.

As my next project, I want to make a similar circuit (ie Single-ended pentode in Triode mode) but using GU-50 tubes as the output tubes for more power.

Has anyone made a similar set-up with GU-50 in triode mode? Assuming I want to use a B+ of around 700-800 Volts, what would be a suitable bias current be? Any suggestions for cathode resistor values and suitable OPT primary impedances to use? Any role for LED biasing as in the Red Light District (RLD)? I will most likely use the James 6123HS OPT as this is an affordable and acceptable quality OPT in my region.

Thanks in advance.

Sid
 
Hi!
When I clicked on this topic in order to help you I discovered that ErikdeBest already answered your question. Curves on that page are real, and you can choose your options from them.
What I can add, it is better to use cathode bias since the amp is SE. You will need more than 100V peak to peak from your driver, so a single 12AT7 stage will be too weak to amplify and deliver such a drive.
 
Hi!
Thank you for the replies.

@erik: Indeed I have no clue about the RH84 design. :-( I am totally a newbie. I just followed the schematic and soldered the thing together. I am not an electrical engineer at all. Maybe you could explain? I got lost in all the jargon about the RH84.


@wavebourn: I am thinking of cathode bias for the GU-50. But I am not sure how much cathode bias ie cathode resistor value it should be if I applied a higher B+ of around 700V.
As for the driver tube, what would be a suitable driver tube? Can I use the 6P14P as the driver tube to achieve the voltage swing?

Sorry for so many questions.

Sid
 
You can't apply 700V in triode connection. Please look at that graphs on top of the page Erik provided, if have some questions first of all read some good book about vacuum tubes and how to draw load lines, and what they mean. There are plenty of such books.
Graphs already suggest several versions of regimes: B+, load resistance, idle current, first grid bias. It is very easy to calculate cathode bias resistor's value: just divide bias voltage by idle current. But before that you must understand tubes and load lines, so looking at that graphs you can understand tube's character, how to ask it to provide what you want.

And of course, you must remember: such voltages and power capabilities can kill!
 
Hi Wavebourn,

You are one of my most respected forumers :) The reason I am trying the GU-50 amp in preference to the other more "common" tubes is because of reading about all your experience with it. It seems like a "super tube", very well made and works extremely well.

As for reading about tubes and loadlines, I have tried to read them but end up very confused. I am not trained in engineering except for basic electronics (and that is solid state electronics) while in school. I will give it another try again.

The reason I am thinking of using B+ of 750V is because my recycled power transformers (2 pieces, one for each channel)) provides AC 750V at 60mA (no centre tap). After rectification and losses through rectifier and choke will probably end up around 750V DC too. What would your suggestion be? Get new transformers? Use in pentode mode? Regulate the voltage to something more manageable? I am well aware of the deadly high voltages.
 
Not many people use GU-50 which is a great thing as it keeps the price low on ebay. Maybe due to the high voltage needed to make them sing, and high power output, and the sockets aren't cheap either.

The schematics I have found for GU50 single-ended have a 250V regulator on the screen, so they are running pentode.

Try valvewizard for theory that does not turn your brain inside out. Its very practical-orientated.
Do a search on "valvewizard"

Edit: Its easy to build a design that is known to work, but when you want to do your own design you need to know the basics behind it, which isn't that difficult to learn but it requires effort.
 
Last edited:
Could I have a look at the schematics in single ended? There are not many GU-50 amps schematics out there and even if there are, many are in P-P mode.


Not many people use GU-50 which is a great thing as it keeps the price low on ebay. Maybe due to the high voltage needed to make them sing, and high power output, and the sockets aren't cheap either.

The schematics I have found for GU50 single-ended have a 250V regulator on the screen, so they are running pentode.

Try valvewizard for theory that does not turn your brain inside out. Its very practical-orientated.
Do a search on "valvewizard"

Edit: Its easy to build a design that is known to work, but when you want to do your own design you need to know the basics behind it, which isn't that difficult to learn but it requires effort.
 
Just a further point: 700VDC + newbie = very dangerous situation

Take note from those who have been zapped, melted clip leads and scope probes, and generally made all the common mistakes at 400V: death is still permanent and one little mistake at 700V will kill you.
 
Hi Wavebourn,

You are one of my most respected forumers :) The reason I am trying the GU-50 amp in preference to the other more "common" tubes is because of reading about all your experience with it. It seems like a "super tube", very well made and works extremely well.

Yes, if to judge fairly one GU-50 must cost at least 5 of KT-88, but both tubes were made for high power and high voltages that is dangerous.
Let's discuss it theoretically; I am not suggesting you to play with high voltages until you have enough of experience.

Here is a picture from Erik's link:


GU50+350v100ma.jpg


Curves on it represent dependence of anode current on voltage between anode and cathode. Each curve was taken by measurement, when negative voltage on control grid was fixed. Different voltages on control grid resulted in different curves.

Blue vertical and horizontal lines represent selected idle point. As you may see, it is 350V and 100 mA. 350Vx100mA=35W dissipation by anode. It is below max rate 40W (actually, 100W shortly is Ok for this tube; in transmitters anodes sometimes were red when antenna detuned), so in such mode the tube will live long life. As you may see, the point where this lines cross each other is right in the middle between -40V and -60V curves. That means, it corresponds to -50V bias voltage. -50V/100mA = 500 Ohm, so 510 Ohm resistor will be fine. Power dissipated on the resistor will be 500x100=5W, so 7.5 - 10 W resistor would be fine. I use in such cases Caddock film resistors, they look like power transistors in TO-220 case.

Pink, yellow and orange lines are load lines for different load resistances (you can understand it thinking of Ohm law: how will look on the graph dependence of current on voltage when resistance is fixed) for 5K, 3.5K, and 2.5K loads. For each load you can draw a table, of dependence of anode current and voltage on control grid voltage.
For example, for 2k5 when input voltage is zero max current and min voltage on anode corresponds to 150V and 175 mA.
Max voltage and min current would mean max swing to the opposite side. 350V (idle) - 150V (max current) = 200V swing, so an opposite swing will be 350V (idle) +200v (swing) = 550V (max voltage on anode).
This max voltage as you may see occurs when voltage applied to the control grid is something like -120V according to graphs.
Now, to calculate output RMS power we need to know RMS voltage from peak one. 200/1.414=141V approximately. 141 squared is 20,000 approximately. 20.2000 divided by 2k5=8W output power. It is quite a big power for SE amp!

What can you conclude? B+? Cathode bias resistor? Output transformer? Voltage swing needed to drive the stage?

You have all data in a nutshell, thanks to Erik!
 
Last edited:
Waveburn I think 8W is rather optimistic. The Pout for 350V/100mA and 2K5 load (without considering the insertion loss of the output transformer) is a bit less than 6W. For 100V peak-to-peak swing you get (from the curves) approx. 360V PP and 130 mA PP which in turn are 127 V RMS and 46 mA RMS. So Pout = 127x0.046 = 5.8W+

@Erik I have had a look at your blog and I can see the same thing about the 45. 1.3W with 202V anode supply and -33V bias is impossible if you don't drive the tube into positive grid. The Pout for this operative conditions is around 0.7 W. I am also sure about this because of direct experience on new/NOS 45 tubes (which test/measure perfectly).

45
 
Last edited:
Thank you 45;
however real life is a bit more complex than all models. Non - linearity, grid currents, losses in transformer, etc..., all plays own role, but for now let's our new friend understand basics.

10K load will give much less power, but with high sensitive speakers sound will be crisp and smooth, like God's kissing.
 
Last edited:
Thank you 45;
however real life is a bit more complex than all models. Non - linearity, grid currents, losses in transformer, etc..., all plays own role, but for now let's our new friend understand basics.
Of course it's more complex and usually is worse. However that estimate is rather good starting point.

10K load will give much less power, but with high sensitive speakers sound will be crisp and smooth, like God's kissing.

Well, I am not sure. More likely you will end up with 2.5W! I cannot see the advantage of this.
Better to go with a 2A3 (a nice chinese Golden Dragon or a vintage 6C4C are just fine) and 5K load then.
With this GU50 I would rather go for lower current at approx. -60V to -65V and use feedback and have more power. Actually I would use it as pentode and take advantage of its higher efficiency. IMHO, of course.

45
 
Hi Wavebourn,

In post #5 you say: "it is better to use cathode bias since the amp is SE".
Please explain (unless you presume a high value grid resistor).


I presume variations of mains voltage and tube tube emission. Cathode bias is a negative feedback by DC current that keeps the tube in the regime. Fixed bias is a workaround for P-P stages that draw current modulated by signal's envelope. If the amp works in class A it is not needed, so you can take an advantage of cathode bias that you loose when have to apply fixed bias to class AB amps.
 
If the amp works in class A it is not needed, so you can take an advantage of cathode bias that you loose when have to apply fixed bias to class AB amps.

Sorry Wavebourn to quote you again but I cannot resist to comment this! :)
If the amp works in class A a fixed bias is not needed only for a textbook example. You yourself have already mentioned a series of issues that actually make fixed bias more desirable in most situations....
If then we talk about the sound is even more desirable.
Cathode bias is just simpler and more suitable for a novice. :)

45
 
What I meant to say is that SE does not include the use of cathode bias.
Think of interstage or cathode follower driven SE class A2 power stages (845 and many transmitting tubes) where the use of fixed bias is normal.
Variations of main voltage is not so much a problem in well designed fixed bias power stages; when main voltage rises so will the negative bias, resulting in higher plate voltage and lower plate current at the same time so the tube will stay in the regime.

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