• 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.

How to make balanced outputs in a pre amplifier

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

I have built this pre amplifier based on the EF86 tube:

http://www.tubelmmp.home.sapo.pt

It sounds well but has limited bandwith. Simulation says that it goes up to 100KHz without difficulty, but low end is quite short (80Hz at -1db) and I am thinking about the caps C15 and C19: I will try to put there about 22uF and see what hapens.
If you have some suggestions for improvement it would be nice.

But now the first question I had: how can I have balanced outputs here? Is there any way of doing this without having to dismantle everything?

Thanks.

Miguel
 
Hi Miguel,

I didn’t realize, does it have limited bandwidth subjectively or you measured it (real roll-off)? If it is just subjectively, some play with coupling caps have sense. In my triode mode EF86 preamp polystyrene coupling caps open the sound.

Anyway, I think it is not necessary to use this circuit for preamp. If I realize, you have attenuation at first stage, then gain of some 50 at second stage. You can make a one stage, triode mode EF86 preamp.

Pedja
 
Hi,

Thank you all for you replies.

I have no test equipment to make objective measurements on it :( . So my opinions are based on simulation and subjective listening.

I have played around C10 and used 0.3 uF on there and heard no diferences on the sound, playing music with a lot of bass (Celine Dion "Ballet", that has really deep bass in the middle of the song). But in the meanwhile I noted that this problem is less apparent, and this is real because I can always compare it with my integrated amp. So some "burning in" of the components must have occurred.

Pedja: I do have attenuation on the first stage, but I already tried it in triode mode (grid connected to cathode), without attenuation, and found the gain way too low. Can you send me a schematic of that 1 stage pre?

Regards

Miguel
 
miguel2 said:
Hi,

Thank you all for you replies.

I have no test equipment to make objective measurements on it :( . So my opinions are based on simulation and subjective listening.

Miguel

Miguel,

Of course good sound must be the final result, but you must make some kind of measuerments, otherwise you might get lost.

You can make the basics at almost no cost:

Audio generator:
Use your PC with NCH Tone Generator: www.nch.com.au/tonegen

Buy or borrow a digital multimeter with AC volts range.
With a sinewave generated, measure the output from your soundcard with the meter to make some calibration data.

Use 1KHz as your reference, then measure the voltage levels at other frequencies.

It doesn't matter if the soundcard isn't perfect, or that the meter won't be. You can even make a speadsheet with the correction coefficients built-in.

With the above, you will be able to measure frequency response to within 1dB or better.

For a noticeable bass loss that you reported, I would guess there must be at least a 3dB loss.

NB For converting a voltage ratio to dB we use 20Log (V1/V2).

With the above set-up, there are many other measurements possible.

For convenience, or if your soundcard is very poor, you could "burn" a CDR from the NCH Generator WAV files, and use a CD player as your source - better than most soundcards.

It is probably best to generate tone at -10dB with the generator for these tests. (That's 10dB below max.)

Good luck
 
Miguel,

I’m not quite sure about gain you achieved with this circuit. With 1000K input resistor overall gain of your preamp is not really big (I checked your schematic in the same program you use), and hence easily achievable with one EF86 in triode mode. One solution you have at

http://www.svetlana.com/graphics/TB/no.23fig3.jpg

There are some differences in circuit I use, and regarding the gain, difference is that I don’t use a cathode cap, so depending on resistors values gain can be 8-12. BTW, this way it will have wider bandwidth. But don’t expect it will help with high-freq roll-off. For this problem, I really recommend trying a few caps, even high-end (film&foil) polyprop (and paper in oil also) can sound murky in this application.

Pedja
 
Hi Pedja and dhaen,

Thanks a lot for your input.

I have downloaded the NHC tonegenerator and it works great. I used it with an AC multimeter and with Winscope, to see wave forms. With Winscope it is not accurate but it will do for a gross estimate. I also suspect of my cheap sound card...

So the frequency response is at -3db at about 50Hz, as you said dhaen. It is in the word file attached, curve with 22uF caps in the cathode. Then I took the caps off and saw that low frequency is good but it cuts at about 9000 Hz, which is not so good. With 0.1 uF film caps gives some nasty effects at the extremes.

So I followed Pedja's idea and put the first stage in triode mode, no feedback. In this way I had not enough gain, so I played with a second stage in cathode follower mode. The circuit is on the same file. It gave the same results as the first circuit with 22uF, as you can see in frequency graph. It also inverts the phase. I have to think a bit more, to go down in frequency. By the way, these graphs clearly show that I am having a hum problem, that I already knew because when I have the volume half way up it starts to do that terrible noise. But with a small crouded box and a 130VA transformer there I just have to think it carefully.

Regards.

Miguel
 

Attachments

  • doc1.zip
    44.8 KB · Views: 63
Hi Miguel,

You can map the soundard's respose with your AC voltmter.

If you want to mess with cathode bypass caps, you must use a large enough value to bypass the entire audio range, otherwise you will get ringing and frequency response scew.

I still think that the output coupling cap (c10 now) is too low.
Also, what is connected to the output while you are making the measurements, and what is the total load impedance?
The AC voltmeter and winscope might be quite low (<1M ohm), in which case it will make the frequency scew seem worse.

As far as the instability is concerned, you don't tell us where the volume control is situated. Please advise.

JD
 
Hi Miguel,

What is the purpose of series resistor at input (R32 – 1000K)?

If you remove it, you’ll get enough gain with one, triode mode stage, even if you lower anode resistor. That means you can remove your current second stage (it actualy doesn’t have a voltage gain, but protects gain of the first stage). Voltage gain of such configuration is about 8x without, and about 20x with cathode cap. Again, why do you use that resistor?

About coupling cap… yes, I also don’t use big coupling cap, mine expirience with this tube says it is possible to use even 68nF without audible cut-off in bass. It actualy do exists, but it is not a lot, -4dB at 20Hz, –2dB at 30Hz, -1.3dB at 40Hz – all at 68K load and 47K anode resistor.

Non-linearities in freq response above 1.5KHz produced your soundboard rather than your preamp.

Pedja
 
Hi all,

Thank you for your replies.

I have to say that in the meanwhile I tried triode mode again and this time I connected the second grid to anode (instead of the cathode). this gave a lot more gain and now I have it that way.

I have measured the sound card response dhaen and it goes fairly well to 20-25 Hz. But somehow yesterday I managed to put about 5 V AC into it and now the computer does not recognise it. I will try to re-install software and hope I have not burned it...
In the tirode mode I used 1000uF caps, as in Svetlana site, and it gives more gain (I measured 27 db) and increases frequency response when compared with no cap. The response now goes to 20Hz at -2 db, which I think it ok. This rises the question about the output cap. I have changed it a few times and saw no changes.
Pedja, the input resistor is necessary, or I get overload from the cd player. I can try to reduce the 100K there, linked to earth, but this is just a voltage divider and I see no advantage in decreasing the 2 values. I do have about 10k to ground in the output of the cd player and this can affect the input impedance of the preamp.

In all measurements the volume control was in the maximum. It is a 100K pot.

I simulated now a new stage based on a 12AU7 but I got some bad results because the outputs v(9) and v(11) are not equal. The first stage is equal tyo what I have now. The first part of the curve goes to 6.4V and the second to 6.6V.

Gromanwe, I use Microcap student version for the drawings. You can download it from their website at http://www.spectrum-soft.com. It is quite easy to start working with it.

Miguel
 

Attachments

  • triode balanced.zip
    80.5 KB · Views: 76
(Hmmm) I use this tube with suppressor grid connected to anode, but I think it has almost the same gain if it is connected to cathode. And that configuration, without any input resistor, never clipped with any CD player with normal output level.

Groman, nice to see your posts again. MicroCap is the program that Daniel Trudeau suggested a few monts ago.

Pedja
 
I do not get the idea, quite???

First you have signal.
For CD output, it should have max 2Vrms.
Most recorded material doesn't have more than 1-1.4 Vrms.

Then you reduce that voltage by resistor divider.
After that you raise the signal again.

Any unnecessary losses in signal's magnitude
should be avoided.
This is because not to have to use as much amplification.

When you amplify a signal with active components,
you always lose something.
You can lose little or more, but the signal is not that good
as it is without amplifier.

So take Pedja's and my advice.
Do not lower the signal before amplifier. You can try to lower the gain in amplifier instead.
To attenuate signal, you use volym control.
That is the only voltage divider you need.

groman
likes amplifiers
but avoids unnessesary amplifying
 
Re: I do not get the idea, quite???

gromanswe said:
First you have signal.
------------------------------------------------
So take Pedja's and my advice.
Do not lower the signal before amplifier. You can try to lower the gain in amplifier instead.
To attenuate signal, you use volym control.
That is the only voltage divider you need.

groman
likes amplifiers
but avoids unnessesary amplifying
Still True! ;)
 
I have this circuit in triode mode for several days and it is working fine. And now I don't need the input resistor :cool: . I also put the volume control before the tube amp and not after, as I had it berfore. It only has more noise but this is not a big problem.
I would like to ask if somebody has sugestions for puting the output balanced. As now I have space for one more stage, I believe I can do it, but just don't know how.

Circuit is:

http://www.svetlana.com/graphics/TB/no.23fig3.jpg

Miguel
 
Hello Miguel,

I’m glad to hear it works OK now. And yes, pot should be at the input. But keep input impedance of the power amp as high as possible, anyway over the 60K. And if you don’t need all that gain you have now, you can find it sounds better without cathode bypass cap.

I did not try to go balanced, but you can find something about it at
http://www.tubecad.com/march99/page4.html and
http://www.tubecad.com/april99/page5.html

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