The Rude Tube

[bernbrue]

Hi,
treble roll off fixed! It was R9 at the ECC88 that wasn´t resoldered correctly. For running the EF86 in triode mode I removed C3 & R3 and tied grid 2 to anode. Cathode to grid 3 connection is left as it is. My assumption that the EF86 could handle line level more easily was wrong. I fed the EF86 with a few demo songs from my Roland JV1010, but the synth didn´t give out full line level. I did some measurements with RMAA. The unit is  linear when the gain pot in front of the 6AS6 is turned almost completely ccw and the output pot almost completely cw. I/O transformers are Sowter 3528 (1:1,5) and Edcor XS 1100 (2:1). Going to update the schematic soon.

I noticed, that the bias pot does´nt have that much effect on the sound with the load line shown in the schematic. I remember that the bias trick worked much better with a different load line of the 6AS6. I would like to have more current going through the cathode thus pronouncing the bias change and the peak modes. This would be the last change and I would really appreciate when someone helped me with the math behind it.
kind regards
Bernd

 

[bernbrue]

Hi,
since I resoldered R9, do I need another coupling cap before R8? HT from the ECC88 goes through R9 and R8 to the anode of 6AS6. We don´t want that, do we?
regards
Bernd

 

[bernbrue]

A few observations with current setup:

EF86 in triode mode results in less gain thus less distortion from the 6AS6. There is still enough distortion, but it hasn´t got this overblasting character. I intend to rewire the EF86 as pentode.

The additional decoupling cap before the ECC88 is not really needed.

I´ll try to implement a suitable input pad before the input transformer in order to avoid the EF86 to distort.

The bias trick is not working as it should. Any suggestions concerning 6AS6 anode load line are very much appreciated.

Peak mode settings work fine with moderate bias and input gain settings. When the 6AS6 is totally overdriven, the effect is less obvious.

regards
Bernd 

 

[bernbrue]

Topic: The Rude Tube  (Read 2006 times), anybody out there ? 

 

[Hank Dussen]

I follow this threat very closely as I find it very interesting but am not experienced enough to contribute.
I'm sure there are other like me. Please keep it up!!

 

[bernbrue]

I follow this threat very closely as I find it very interesting but am not experienced enough to contribute.
I'm sure there are other like me. Please keep it up!!

Glad to see that there is a sign of life! I haven´t got that much experience either, I learned just by trying. I still haven´t got an electric shock. Keep fingers crossed.  ;D
regards
Bernd

 

[kepeb]

agree with hank.
following closely.
I have almost offered my thoughts on certain things while following but am too inexperienced to offer any expert advice. most i can do is bump your thread again when it goes quiet
for instance, i would of thought a cap between r8/c4 and r9 would have been a good idea but i cant back that up with much but assumption of what the circuit is doing.
and i also would have left in the cap on the DI for my own piece of mind even tho its probably superfluous. (it is possible to put dc in here by accident and also makes it safer imo)

I'd like to have a go at making your 'rude tube' but will need some help with the power supply and the regulated high voltage.

not sure what to suggest with the bias...

 

[bernbrue]

It´s not that difficult. If you want to built it PtP have a look at the PSU of the DAOC-Compressor that handles 3 tubes per channel quite easily. In the "Redd P2P thread" there are also a lot of details explained how the preamp works.  Otherwise check Guavetone´s Orange 86 version of the preamp (boards available), built it, remove the NFB loop (basicly leave out the gain switch), lift R8 and and insert the 6AS6 valve stage as shown in the schematic. I would really love to see someone else building the "Rude Tube" in order to share experience.
regards
Bernd

 

[Kingston]

Can you reiterate the problem with the bias? I have a hard time following what the problem there is. Did it used to work? and then you changed what?

 

[bernbrue]

Hi,
thanks for moving the thread to the Lab. Well, there is not really a problem with the bias, Maybe we should not worry about this too much and fix (improve) it later.

Yesterday I played around with the prototype and realized, that the output pot behind the ECC88 changes the frequency response significantly (lowering output gain results in loss of high and low frequency). So there are two solutions: Implying the pot before the ECC88 or as Kingston suggested a 600 Ohm T Pad attenuator after the output transformer. The latter solution needs a transformer, that can handle the huge amount of gain without distortion. I´ll try to test the first idea and report back.

regards
Bernd

 

[bernbrue]

Oh yeah, that´s much better. I installed decoupling cap and output pot right after the 6AS6 stage, so that it controls the signal going to the ECC88.
Direct A/B comparison between signal going to the mixer and the signal feeding "The Rude Tube" in triode mode with unity gain are very promising. Fundamental bass response and very crisp highs.  Further measurements and schematic update will follow.
regards
Bernd

 

[bernbrue]

schematic updated (1st post)
regards
Bernd

 

[bernbrue]

The unit sounds excellent now. I really like what it does in triode mode. It could be described as fattening or sweetening, if you like. Fine so far, but now it gets difficult. Back to the bias "problem". I don´t know the technical terms for it, but I would like to bias the 6AS6 closer to the cut-off (cold biasing?) in order to produce a ‘harder’, more ‘crunchy’ sound only by changing the bias. The distortion I get now is just achieved by overdriving the 6AS6. I think it´s a kind of clipping on one side of the waveform. I need more bias voltage and less anode voltage. So I have to change the load line and take anode resistors with higher values. In short the tube´s working point is not optimal. Lowering anode voltage would also result in less gain, which is an advantage in this design, since there is almost too much gain coming from the 6AS6. I need help from the tube gurus.
regards
Bernd

 

[guavatone]

bern, did you draw a loadline and go through the basics of setting a bias point?

 

[bernbrue]

not really, I took the anode resistors as seen in other circuits. I changed the values quite often to find an optimal working point but trial and error doesn´t always lead to good results. I´ve read a lot about calculating load lines, but musicians and math aren´t really good friends  ;D
regards
Bernd

 

[guavatone]

Arg. The curves on this one are bizarre too.  Hmm.  I can try to take a look but it may take some time.

Most curves slope up.  this baby is funky:
http://www.mif.pg.gda.pl/homepages/frank/sheets/009/5/5725.pdf

 

[bernbrue]

Yeah, really funky. I´ve studied all these curves in the data sheet and I didn´t manage to calculate a reasonable value for the anode resistors. I´ve read somewhere that the small signal pentodes like to see fairly large anode resistors. I realized in a former setup that the tube behaves nice in the 60V area.
regards
Bernd

 

[guavatone]

I would try 3 possible load lines:

1.  With the Supple at 170VDC try to bias Anode at 7mA with a 24K Ra.  Bias the Kathode to -2V with Rk at 280R

2.  Use Ra=16K and bias K at -1.5V with Rk=240R


3.  Bring the voltage down by raising the 50K resistor to get 150V.....  I will try to calculate values later.


It looks like it's a comparably low volt high current tube.  And running at a lower current with the large 47K Ra will not allow the signal to swing, because it wants to be biased at the -3V Kathode bias

 

[guavatone]

Also check this out:

http://www.valvewizard.co.uk/
click on pentodes

The load lone should go through the knee of the curve and hit the curve somewhat halfway between 0V and 170V on the bottom line.

I have been looking at triode curves lately and forgot that pentodes are curved like this.  Why did you choose this input tube?

 

[guavatone]

Now for the screen calculation:
Plate to screen(g2) ratio should be 5.2/3.5 or close to 1.5.  According to merlin's calcs, the screen resistor should be about 85k, try 91K here, in scenario #1 wit a 24K Ra.

OK try this:

change 139K resistor to 91K
change 47K resistor to 24K
change the 1K5 resistor to 280R and use a 100uF 16V cap to bypass the low end.
-I recalculated with screen at 2.8Amps and operating point at 4.2Amps

If any gurus see this and I am totally wrong, please do tell ;-)

phew....I need more excel calcs to make this easier