REDDI DI

[ruffrecords]

ok playing around with this 6DJ8,

now i am wondering if the schematic needs correcting, that 6N1P is shown with separate 220 ohm cathode resistors, but since the DJ8 cathode resistor turned out to be 220 ohms for both, and the REDD 47 has a 220 ohm resistor for both cathodes also, maybe the schematic should have both cathodes tied together, always wondered why they went through the trouble to use add 2 more parts with no apparent reason,

Hi CJ,

The reason you see differences is because the 6N1P is not an equivalent of the 6DJ8 despite many web sites saying they are. They have the same mu but the gm and ra are quite different which means the biasing conditions are different. The 6N1P also has a plate rating of250V, much higher then the 6DJ8.

Cheers

Ian

 

[CJ]

thanks Ian!  i read that in your earlier posts and it was in the back of my mind as i was biasing the DJ8,

i still wonder about the dual cathode resistors, maybe Garry traced both cathodes back to the same cap and resistor, maybe not if the bias resistors are different by a factor of 2,

here is a graph of the DIY attempt of the OPT,

inductance vs freq vs air gap,

i like 3 mils, looks the most linear but still has plenty of Henries at 5 hz,

we do not care about the gap after about 3000 hz, as the core is no longer the dominant factor in freq response, capacitance will be the main issue, starting at about 5 or 10 k hz, plus, all gaps converge at around 1 k hz, so we limit the graph to get better resolution,

 

[ruffrecords]

Well, CJ, it's meant to be a bass guitar di box so we need to take good care of the bottom end. My vote goes for 5mils.

Cheers

Ian

 

[CJ]

10 mils might be a straight line at 50 H till 100 hz if you want real linear response,

i will plot that after my 20,000 winks,

 

[CJ]

ok low ratio xfmr means we can get flat response without a million sections,

i bet we can get flat with just sec-pri-sec if we play around with some winding tricks,

enclose the pri with the sec's to get good linkage between the pri and sec,

use reverse winding on one sec to balance the C,

should be good to 60K just like the original,

834T-5200T-834T-reverse  should do the trick, will wind one up and see,

 

[CJ]

DIY OPT for this box now has 4500 T pri and 1420 T Sec, Sec-Pri-Sec

forgot to reverse wind the outer Sec,

so we can do a cool experiment by checking freq response both ways,

going to rewind that section after i do a freq plot the way it is,

here is how the inductance came out with various gaps,

5 mils seems to be the ticket, 10 mils and the core starts to sing when the I bars are not clamped down,

there is a confusing equation in the transformer book that says

u1Q1=u2Q2,  that is, as perm goes down, the Q goes up,

but Q is Quality Factor, right? XL/DCR ?

so if DCR is fixed, and perm drops, so does inductance and therefore XL,

so how can Q go up as inductance drops? this equation must be wrong, 

no, there are 2 types of Q, Quality Factor and Bandwidth,

so the equation really says that as perm goes down, selectivity goes up, that is, you have a steeper curve, so Q depends on DCR and Perm.

you can see this in the graph below, as we increase the air gap, perm drops, and the curve becomes steeper=more selectivity,

also note that the bigger gap shifts the peak to the right,

weird that the inductance paek shows up after we add the secondary winding,  ???

 

[CJ]

wired up this box, sounds great, but the 6DJ8 is being a pain even with 150 volts on the plate, kind of unstable,

maybe a 12AU7a would be a cheap alternative, without the pot on the output, this DI was lighting up the overload LED on the yamaha mixer,

with the pot you do get nice tone variations for bass, crank it way down and you get a nice deep tone,

the DIY transformer sounds great, settled on a 3 mil gap for better response below 20 hz,

use the 1K pot, 10K needs to be cranked down most of the way, even with the audio taper,

hopefully i can find a quiet tube, otherwise i like the country hick for a DI the fets are very quiet,

 

[dandeurloo]

The Reddi I used was quite.  I wonder if we missed something or if they select tubes by and for them?

 

[ruffrecords]

The Reddi I used was quite.  I wonder if we missed something or if they select tubes by and for them?

The 6DJ8 is generally a quiet tube but they have a high gm and can be prone to oscillation. A pair in parallel wil have an even higher gm so a 1K grid stopper wired right next to each grid pin mi cure it. VHF oscillation does often sound like noise in the audio band.

The 6N1P used in the REDD DI is a completely different tube; it has a lower gm so it will be more stable and a higher ra so its gain will be a lot less. It is also known to distort a lot more than a 6DJ8.

Cheers

Ian

 

[My3gger]

I agree with you Ian, 6N1P is very different tube and no mater how you bias it, it will never sound like ECC88. This differences for sure influence DI's sound which seems to be well thought.

 

[CJ]

what is the best brand of that 6N1P or is it just one company now days that makes it?

this DJ8 is not that noisy unless you crank up the volume then you get the typical plate hiss,

it does sound great for guitar also,

the tube i used was a "pull" so i am going to get something new from the tube caddy,

here is a blurb that spells it out>

"
The generic tube of 6N1P is  ECC88, which was originally designed as a medium
voltage tube (90Volt) for measurement equipment, with very high linearity and low
distortion, such as required for oscilloscopes. Today, ECC88 is very popular as an
audio tube, since output impedance is relatively low and gain is relatively high.
However there is a trend to use ECC88 at highest possible voltage, and highest
possible filament to cathode voltage, and highest possible specifications in general,
and expect no noise and excellent lifetime. Because of the limit breaking use,
ECC88 is sometimes not good sounding, and/or develops a noise, or short circuits
inside with sparks. It must be said, the gentlemen circuit designers are
the real source of the trouble. It is accepted (by most) that the maximum limits should not be exceeded (also not a little bit) but it is at the same time expected by most, that if you stay a fraction below maximum limits, the tube should perform fully
normal, and have no reliability issues, and all tubes out of a batch have “typical” specifications. This is in strong contradiction about what (real) specialists and historical literature tell you on the subject, but I have given up on the idea of learning somebody something about this issue.
Also with the issue if typical specifications, we see another classical error. “Typical”
means many tubes from different batches will have this value as an average. When
minimum and maximum values are not specified, it means you must expect +/-
40% for USA tubes and +/- 30% for West European tubes. A good circuit designer will
select such a +40% and –40% tube and will make sure his circuits work good on
both. Beginners take random tubes, and design the circuit around those.
Anyway, in a few words, here is what causes the disappointment with ECC88. If you
have trouble with an ECC88, changing to 6N1P may be the solution, since you can
not change the circuit.

6N1P is available in three versions.
1.Standard. Which al already much tighter specified as ECC88.
2.6N1PEV (or called EB, that is the same). Long Life.
3.6N1P-IV. A very rare version with very tight gain specification"

 

[CJ]

ok this thing sounds killer with a 5963 tube and an 820 ohm cathode resistor,

both cathodes are tied to the one resistor,

low noise, killer bass,

sounds a bit better than the country hick for bass and even less noise!

so the lid is going on and the box is getting bolted to the rack,

tried a 12AU7 but i like the 5963 a lot better, very quiet,

going for a 1/4 inch output jack as this thing can drive a line amp, no mic pre needed like the country hick,

output trans is sec-pri-sec,

710 turns - 4500 turns - 710 turns on 75 EI with a 3 mil gap, steel lams,

ratio is 3.17:1 = 6K:600

5963 is a premium grade 12au7a,  $6.65 @  Antique Electronic Supply

tube data:

http://www.mif.pg.gda.pl/homepages/frank/sheets/049/5/5963.pdf

 

[ruffrecords]

Unfortunately that blurb is riddled with errors. As I have said before , the 6N1P is not an equivalent to the ECC88. Their only common parameter is mu. This is not simply my opinion. I have measured many samples of both types.

Cheers

Ian

 

[abbey road d enfer]

Unfortunately that blurb is riddled with errors. As I have said before , the 6N1P is not an equivalent to the ECC88. Their only common parameter is mu. This is not simply my opinion. I have measured many samples of both types.

Cheers

Ian

Certainly the transconductance is quite different 5 to 12.5! Makes a hell of a difference.

 

[chrispsound]

Would this Edcor off the shelfer work?

http://www.edcorusa.com/p/475/cxse25-600-5k

Thanks, ChrisP

 

[abbey road d enfer]

Would this Edcor off the shelfer work?

http://www.edcorusa.com/p/475/cxse25-600-5k

Thanks, ChrisP

I'm sure even this little one would work
http://www.edcorusa.com/p/546/gxse5-600-5k
You don't need to go down to 20 Hz for DI.

 

[chrispsound]

Would this Edcor off the shelfer work?

http://www.edcorusa.com/p/475/cxse25-600-5k

Thanks, ChrisP

I'm sure even this little one would work
http://www.edcorusa.com/p/546/gxse5-600-5k
You don't need to go down to 20 Hz for DI.

Even though it is going to be used primarily as a bass DI?

 

[abbey road d enfer]

Would this Edcor off the shelfer work?

http://www.edcorusa.com/p/475/cxse25-600-5k

Thanks, ChrisP

I'm sure even this little one would work
http://www.edcorusa.com/p/546/gxse5-600-5k
You don't need to go down to 20 Hz for DI.

Even though it is going to be used primarily as a bass DI?

4 string bass goes down to 41, 5 string B goes down to 30.5.
But in fact what you hear is the 2nd and 3rd harmonics at 64 and 96 Hz.
Most bass cabinets don't even go down to 50.

 

[azure skies]

and here is a pwr scheme!!

Are you sure about those resistors? The LM317 output voltage is determined by 1.25(1 + R2/R1). Those values would give an output of 2.5V. It seems R2, the resistor to ground, should be about 360 ohms.

 

[dirty1_1garry]

and here is a pwr scheme!!

Are you sure about those resistors? The LM317 output voltage is determined by 1.25(1 + R2/R1). Those values would give an output of 2.5V. It seems R2, the resistor to ground, should be about 360 ohms.

Yeah, resistors around LM317 ia wrong!! That's my wrong read of colour codes
Here is a good LM317 calculator:
http://www.reuk.co.uk/LM317-Voltage-Calculator.htm
It said that for 6.3Vout resistors should be 240R and 970R(to ground).