Overkill tube pre/EQ called Drive-1. Maintenance update.

Help Support GroupDIY:

tv

Well-known member
Joined
Feb 22, 2006
Messages
1,459
Depending on how it is loaded, a cf may rectify some of the ac it passes, so the resulting distortion will be in the very essence different to what is usually "expected" as tube-distortion. In a way similar thing can also happen with solidstate followers..

But paralleling the tube could also be a way to somewhat diminish the need for a cf ..
 

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
tv said:
But paralleling the tube could also be a way to somewhat diminish the need for a cf ..

Just that the paralleling would happen at the very first stage where there's no need for low impedance. It only needs to handle the 100k attenuator.

There was never any need to lower the impedance of the output stage (it's like 1k5 ECC99). I think the only reason plumsolly used CF was to reduce the gain to unity vs. the common cathode stage (20dB or so). For whatever reason it seems the massive combined gain of all the stages is causing oscillation for him.

I my preference of design a CF is only allowed as an unimportant utility stage, like guitar amp FX loop drivers and such. Otherwise it's like wasting a perfectly good tube and certainly "not allowed" at the output where we want "sound". I feel pretty much the same way with any tube design swimming in negative feedback. I mean, might as well go fully solid state at that point. This is entirely an efficiency point of view. We can have near perfect performance cheap with solid state. Attempting anything approaching that kind of performance with tubes is a futile and very inefficient effort. Use a tube today where solid state fails ie. distortion.
 

tv

Well-known member
Joined
Feb 22, 2006
Messages
1,459
If you were thinking on using a "IRF" in place of a cf, they have some (and then some) gate capacitance so this would dictate thinking both in terms of top-end response and dangers of oscillation.

Hi-fi folks have had good times using mosfets as active plate loads, so this may be worth of further experimentation but will definitely alter the "tube tone" that most ppl want to hear from their instrument/mic preamp..

IIRC Analag had some kind of fet-loaded tube pre? Perhaps I'm wrong..

Ps. paralleling in early steges may shave off a couple of dBs noise..
 

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
This maintenance update might be of some interest to everyone working with tubes.

The heater regulator started to glitch and then finally completely failed after some years of active service. It had a massive heatsink and temperatures were always in check. Still this 5amp rated part died. I replaced with dumb resistor. The heaters eat 3.5amps and i dropped to roughly correct voltage with two parallel 25w resistors, around 1ohm with the transformer i am using. I compared specs of a newly built unit with this new "dumb resistor" version that has years of duty under the belt (RMAA measurements). No change with noise floor!

that was officially the last tube heater regulator i ever built. There is simply no point and now i finally have data to back it up. Schematic on the first post of this thread was updated to v2.1 with the regulator removed.

Also, when the regulator failed i wasn't running audio through the unit. I did not know about the problem so all the tubes ate raw b+ for a full day with no heaters. No change in specs (or sound). Another tube myth debunked right there.
 

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
Another thing of note is that while the heater regulator was noiseless (below 0.3mVAC) and the dumb resistor produces a whole 20mVAC there is no change in noise floor. This difference could be easily detected in noise floor if it happened with B+.
 

millzners

Well-known member
Joined
May 7, 2010
Messages
219
Kingston said:
This maintenance update might be of some interest to everyone working with tubes.

The heater regulator started to glitch and then finally completely failed after some years of active service. It had a massive heatsink and temperatures were always in check. Still this 5amp rated part died. I replaced with dumb resistor. The heaters eat 3.5amps and i dropped to roughly correct voltage with two parallel 25w resistors, around 1ohm with the transformer i am using. I compared specs of a newly built unit with this new "dumb resistor" version that has years of duty under the belt (RMAA measurements). No change with noise floor!

that was officially the last tube heater regulator i ever built. There is simply no point and now i finally have data to back it up. Schematic on the first post of this thread was updated to v2.1 with the regulator removed.

Also, when the regulator failed i wasn't running audio through the unit. I did not know about the problem so all the tubes ate raw b+ for a full day with no heaters. No change in specs (or sound). Another tube myth debunked right there.

I had a cold solder joint on one of my heater circuits without me knowing it and ran that way over night.  I was freaked out it would kill my tubes...  Didn't seem to matter. 

So now that you're on board with unregulated DC, do you think you'll ever come around to using AC heaters?
 

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
millzners said:
So now that you're on board with unregulated DC, do you think you'll ever come around to using AC heaters?

I have not been able to get better than -60dBu with ac heaters, but it's a long time since I lasted tested and measured it(was very similar set up as drive-1 and these line level measurements.) It's certainly worth a try and my layouts and wiring has improved since. but i find it somewhat liberating not having to worry about heater wiring and getting more like  -90dBu with ease.
 

gyraf

Well-known member
Staff member
Moderator
Joined
Jun 4, 2004
Messages
10,385
Location
Aarhus, Denmark
You may want to consider about tube longevity - fluctuating mains voltage and high turn-on currents can probably make life hard for the tubes. Not a real problem if you use easily-available new-production tubes, but I think it a shame to inadvertently burn out good NOS's

I have seen considerably longer average tube life after moving into regulated heaters - my guess is main advantage comes from limiting power-up current.

Jakob E.
 

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
gyraf said:
You may want to consider about tube longevity - fluctuating mains voltage and high turn-on currents can probably make life hard for the tubes. Not a real problem if you use easily-available new-production tubes, but I think it a shame to inadvertently burn out good NOS's

Fluctuating mains voltage should be a non-issue. 10% tolerance specification on those NOS tube and very stable infrastructure guarantee this. I suppose your view as a commercial developer differs since you have to cater for someone who needs a varimu compressor in the middle of the Amazon rainforest or close.

With high initial currents you might be right. Any tips on how to deal with these currents closer to, or above 5A? I no longer trust LM338.

Another consideration is I'm using inexpensive CCCP-era tubes. Less than $5 each, usually closer to $2. Maybe I shouldn't overthink this.
 

gyraf

Well-known member
Staff member
Moderator
Joined
Jun 4, 2004
Messages
10,385
Location
Aarhus, Denmark
Kingston said:
With high initial currents you might be right. Any tips on how to deal with these currents closer to, or above 5A? I no longer trust LM338.

Maybe a higher raw voltage, allowing for a bigger series resistor? Maybe PTC or soft-turn-on like in large-toroid poweramps?

Maybe a simple current-source/limiter with a high-power transistor?

The problem is only in the first 15-20 seconds after turn-on

Jakob E.
 

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
gyraf said:
Maybe a simple current-source/limiter with a high-power transistor?

This crossed my mind also. It wouldn't even need to be a very high performance gadget, just something that can take the initial heat and not burn and fail. Soft-turn-on would be an easy additional feature on top of that - perhaps overkill. But what's a PTC?
 

rock soderstrom

⚽️⚽️⚽️⚽️⚽️
Joined
Oct 14, 2009
Messages
1,213
Location
Berlin
Interesting discussion!

Wikipedia:
A positive temperature coefficient (PTC) refers to materials that experience an increase in electrical resistance when their temperature is raised. Materials which have useful engineering applications usually show a relatively rapid increase with temperature, i.e. a higher coefficient. The higher the coefficient, the greater an increase in electrical resistance for a given temperature increase.

I am not sure, but I think a NTC should be the way to go??

 

gyraf

Well-known member
Staff member
Moderator
Joined
Jun 4, 2004
Messages
10,385
Location
Aarhus, Denmark
yup, positive temperature coefficient resistors - the blue disc-shaped thingys you often see in computer smps's for limiting startup currents in primary side. Maybe/Maybe-not useful in a scenario like this - I'm not entirely sure about their timing and thus suitability for longer current spikes.
 

plumsolly

Well-known member
Joined
Nov 3, 2007
Messages
969
Location
Denver, Colorado
I took a break from this to start a new job and move, but I am now back at it. I have the circuit pretty much as drawn right now, and I have done some more probing and found a few interesting things. The source of my ~45Khz oscillation seems to be radiation from V2. Tube shield helps, strategically placed shielding plate in chassis helps, grid stopper on V1B helps - I have something like 65K for the GS right now - I have not swept it to see how much high end I have lost. I have it stable on most settings now and am just trying to get the last few sorted out. I am using a Russian 6N9S for V2, and I will try an American tube to see if it helps. Any other ideas of stuff to look at here?

Thanks,

Ben
 

plumsolly

Well-known member
Joined
Nov 3, 2007
Messages
969
Location
Denver, Colorado
Got it working, oscillation free, through a number of tweaks. Most significant was the addition of the tube shield on V1B, I believe. Now on to testing and the final wiring and panel stuff.

Ben 
 

plumsolly

Well-known member
Joined
Nov 3, 2007
Messages
969
Location
Denver, Colorado
I have it working and all wired up and am now fighting a 60hz hum problem. I have got it narrowed to V2B - if I ground V2B grid, hum goes away - if I ground V2A grid, hum remains. I have r10 and r11 right at the tube, and then I run screened 2 conductor cable to the front panel to connect to VR2 and VR3. Screen is connect to chassis at only one end. I am running AC, elevated heaters. Any thoughts?

Edit: Meant grid above, not cathode.

Thanks,

Ben 

 

MidnightArrakis

Well-known member
Joined
Sep 25, 2015
Messages
134
Here's what I've been working on lately.

At heart it's a "one bottle" 6SN7 preamp with a unity gain treble and bass tone stack between the 6SN7 input and output stages. Then there's a whole separate line amp as the output stage. I have already used this line amp in several other projects and basically don't want to live without it anymore. If you don't need the gain the line amp part can be considered optional, but it will also happily drive a 2:1 output transformer and sounds wonderful when overdriven. The V1B stage alone (6SN7 or variant) used as output will need at least 5:1 ratio.

There are just about no tricks in this, as any tube guy will quickly see. I don't intend to cover any new ground here. I'm just gluing together some proven solutions I like, and as such I don't really deserve any credit for it. Advanced copy and paste with my own preferences for tube bias. This is a draft and I will end up tuning bias, pads and tone stack for best sounding distortion points when I hear them first.

The design is absolutely loaded with features and switches, most meant to be relays to minimise hum when I start laying out the thing. You can downgrade performance in several ways and the unit is basically a template for experimentation with tube distortion. There's a possibility of feedback, no feedback, over-bias, under-bias, cathode bypass, no cathode bypass, starved plate and what else! At optimal settings it will be a turbo-charged clean preamp with plenty of headroom. When all options are switched to "completely wrong" it makes various types of distortion/overdrive. Will work great for line level sources, and that's what I will be mostly using it on. I have intentionally biased the optimal settings a bit on the high current side at the expense of raw gain and voltage swing. Tubes seem to sound better this way.

At unbypassed setting max gain is around 75dB (with no output stage feedback).
Cathode bypassed setting max gain is about 90dB (with no output stage feedback).

PSU is a brute force regulated affair with far too much filtering and voltage dropping just in case. This is supposed to be a very wasteful current eater, I don't care.

If you spot errors that will downgrade the design, or have more ideas, now is a good time to mention about them.

I already have a question on the use of feedback. Do I need to add a cap between C7 and R12/R13? Will the R16 somehow screw up (bias?) the tone stack feedback?

What about the output stage feedback? Do I need a similar cap there, or is C12 (connected to output trafo) just fine already?

Thanks,
Mike

[edit]
revision 1.9 with many fixes and various tweaks. There used to be NFD loop around the last two stages, but it was completely pointless. The design does 0.1-0.3% THD+N even without it - mostly second order harmonics, invisible. The option is still there on the PCB. PCB's have seen various levels of development and prototyping too, and these seem to work well, giving the builder plenty of options for further experimentation. The "plate starve" option of the PSU that drops B+ voltage is of the dumb-but-works school of design. It wastes a stupid amount of watts and needs a big-watt resistor, but whatever. The EQ frequency selectors, including the caps are probably best wired directly on rotary switches. The relays limit the choices to only two frequencies per bass and treble control. nF and pF range caps are tiny so this is easy to do.

initial working and well tested release 26.2.2012

maintenance update 3.6.2013
Note, the above overdriving_amplifier_PCB_1.1 is not perfectly compatible with the below update. PSU was not changed. see this post,


maintenance update 21.1.2014
Heater Regulator removed. see this post,






For reference here are some relevant threads that led to this design.
Kingston: Is your schematic version -- http://www.michaelkingston.fi/files/Drive-1_schematic_rev2.1.pdf -- still current and valid??? Or, have you implemented any changes and/or updates since then? If so, what??? And, do you have a more current version of said schematic?

Since I am both a "Senior PCB Designer" >> AND << a "Senior Mechanical Designer", whenever you become involved in any other similar projects such as this "Drive-1" project of yours, let me know and I will do what I can to help you out, OK???

NOTE: My Grandmother was born in Helsinki, Finland and came to the U.S. when she was 7-years old sometime shortly after 1900!!! She was a very special lady and she had such a special charm about her that I can clearly remember!!! I have dreamed about visiting Finland just to see how life is over there!!!

/
 

Attachments

  • JBW-Designed - RACK STUFF & PCB's.pdf
    9.6 MB · Views: 23

Kingston

Well-known member
Joined
Nov 1, 2005
Messages
3,709
Location
Helsinki, Finland
Is your schematic version -- http://www.michaelkingston.fi/files/Drive-1_schematic_rev2.1.pdf -- still current and valid??? Or, have you implemented any changes and/or updates since then? If so, what??? And, do you have a more current version of said schematic?

This is it. The schematic is perfected - not a single fault ever in that latest revision. It has been performing brilliant, in fact sitting in the master bus with music running through it right now. The tube selection is also such that I don't expect having to open up the unit in several decades.

Since you are a Senior PCB Designer you can of course tell the PCB is amateur quality. It isn't fit for paint by the numbers projects at all, nor could it be sent to modern PCB facility for large batch printing. But those fat traces and large ground planes make it really easy to work with. It was always perfectly noiseless, as in the ground design is good. Similarly, those large format tubes needed decent mechanical support, and that's also catered for. If I need to build a second unit, I would not hesitate to use these again.

Thanks a lot for the offer to help. If you had been around here a decade a go, I can easily see how we could have done this together. Currently I am far too busy elsewhere to get into new designs. Also the racks are full!
 
Last edited:

Latest posts

Top