• 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.

6V6 line preamp

Since I replied to another post about the 12AX,12AU and 12 AT series of tubes I mentioned that I didn't like the 12AU7 sound too much. However I tried it only as a gain stage with gain wide open. So I felt that I should check it with negative feedback to reduce gain as it was far too much in my application.
I just modified my older board and did just that so that I had only about 14dB gain . I also changed the electrolytic output cap to the same film cap.

It sounds very good now. In fact HF seems clearer than the 6V6. However the difference isn't 'large' . Just noticeable on some tracks. Mainly nitpicking differences. Both preamps now sound very good. Main thing is the 12AT7 isn't microphonic and 'looks' less impressive ! :) The tube is a Mullard NOS tube.

Bass is slightly different in it's SQ not so much in quantity. Need to do a longer listening test with equalised levels. But I'm glad I tried this. I didn't have the right resistors but used something close. I was concerned about the Zin from dropping too much. Hum is less than the 6V6 but still there. Same dc supply ! Reduction must be due to the NFB.

Lots more work to do. I'm giving away one preamp to a friend and so must decide which one it should be. :)
 
diyAudio Chief Moderator
Joined 2002
Paid Member
I had 12AU7 cascade line pre (CC+CF) which was good but tonally bland vs a big bottle trioded single stage. A noval you may like even better also is 12AY7. You probably hear further differences due to different Zo with the feedback now. Loop feedback beyond 6dB was always cleaning up some grunge but also flattening out the soundstage for me on such medium linearity novals. Some hum maybe nests in the rail too. The 6V6 draws much more current through a smaller anode load.
 
Yes, Salas, my setup isn't ideal. Transformer very close to the signal cables etc. I will need to clear up everything first and keep only the preamp on the work table. I haven't used the scope to check anything yet though it's next to the unit. It runs directly on the mains and has no isolation transformer. Sometimes it appears to cause some hum when it's earth lead is connected ! Maybe the mains earth isn't ideal !
 
diyAudio Senior Member
Joined 2002
Hi,

Thanks Frank. That's what I saw long ago.
So does the grid actually draw current on a positive transition of the input signal at low mV levels ? Vgk = 0 of course . There is no coupling cap as usual.

The grid potential is still negative enough since it auto-biases itself. So no grid current is drawn and it won't be driven positive either with the output voltage of most MC being so tiny.
Hence no input cap is required.
The major difficulty nowadays is finding J-fets with the same IDSS as the BF244A. Or finding BF244As.
They're still around but are getting scarce and rather expensive though.

Cheers, ;)
 
Wouldn't the BF245A be sufficient ? That's available and looks similar in Idss and some of the other parameters.

Sometime ago I tried a regular triode gain stage with resistive and CCS ( mosfet I think ) load . The resistive load sims with higher distortion of course but the resistive load sounds better !
Did anyone check this ? Maybe because of the high 2nd order distortion ? More distorted but sounds better !
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
An active CCS load is a horizontal load line when a resistor is a slant load line. In high enough swing there are obvious differences to be had in THD and its profile. Usually there are subjectives involved in choosing that depending on the designer's approach. But in very low swing for very low input signal applications a CCS is put there more for the PSRR benefit it brings. Someone can use a Cascoded JFET for that, in a J113s small number will be easy to have samples between 15-25mA for instance. Or a low Rbb' NPN. DMOS maybe is more noisy in a critical application as this and it would not work well if less than 20mA is needed for a number of starved noval tubes. Also if using individual same current CCSs for each plate relaxes the tubes matching adventure hunt and allows to be focusing in low grid leak current samples primarily. In the end you put the total mousetrap through the high bit FFT where the best audio band noise floor and 1/F behavior combination wins. THD is plenty low in such diminutive swings anyway.
 
Wouldn't the BF245A be sufficient ? That's available and looks similar in Idss and some of the other parameters.

Sometime ago I tried a regular triode gain stage with resistive and CCS ( mosfet I think ) load . The resistive load sims with higher distortion of course but the resistive load sounds better !
Did anyone check this ? Maybe because of the high 2nd order distortion ? More distorted but sounds better !


May be because there were different anode voltages with CCS and a resistor, so load line with a resistor was going through more actually linear regime? I never saw anybody who really liked more distortions, despite of their beliefs.
 
If we have curves going pretty much parallel in the area of operation there is not that big of a difference between a current source forced loadline and a common resistor loaded line. If on the other hand we have a region where the Vgk curves looks more like a sun feather in the choosen region and choose a resistor as load that also goes steep we can see how the non linearity kicks in.
 

Attachments

  • parallell.png
    parallell.png
    14.8 KB · Views: 441
  • sunfeather.png
    sunfeather.png
    19.1 KB · Views: 447
Yes the difference may not have primarily due to the CCS alone. It wasn't a focused test. It was just 'noticed' while doing many other things. Could have been due to something else. However I think I'll try it again just to make sure and get it off my mind. Thanks for the comments.
 
Went back from the 12AU7/6992 preamp with NFB to the single 6V6.
Something nice about the 6V6 but I still feel it's a bit 'shouty' when played louder. Must investigate with a FFT. Bass sounds a bit deeper and taut though both circuits use the same type output capacitor. Only one series cap in the signal chain. No cathode resistor bypass capacitor.

Zo isn't the cause as the noval socket tubes have much lower Zout ( uses a cathode follower !).
Am off to bed. Been up very late the past two days !
 
Reduced gain with output to input grid resistor and series input resistor. ( 220K/39K ) . Dropped gain a lot but seems to sound nicer. Tomorrow I will try out the discs I used earlier. Will also have to check with cathode bypass cap.
Unless my mind is playing tricks, this seems better. It's too late now. I'm going to bed.
 
I added the NFB to cut gain . I have too much without it . Bypassing the cathode resistor will reduce Zo further. But this isn't important for me. Sounds good as it is. I tried bypassing but I don't find a significant difference. There is a slight level change due to the change in Zo. My current load is 47 K ohm.
Not much tim eto check minute differences. I haven't fixed the hum problem yet!

Will have to shelve everything in another couple of weeks as I am moving from here.
 
If someone has the inclination to try something new, I might suggest some slightly different operating points that I have arrived at in my system. 350V B+, 15k Rl, 140V +/- Va, 680Rk, Ia = 14mA, Vg = -10V. I have 3.3uF on the output, but did not hear any real difference over 2.2uF. It is driving an amp with a 47k Zin, and there is no issue in terms of the higher output impedance of the preamp. I haev tried 10K as well as Rk setrtings as low as 250, and this seems to be where i settle. Lower Rk brings bias up, but i like the balance at this lower setting. My impression is that it has more punch now and yields better overall presentation, but maybe its just different. It would be nice if someone else gave it a try to see.