What tubes for a tube amp? - Page 7 - diyAudio
Go Back   Home > Forums > Amplifiers > Tubes / Valves

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 19th February 2010, 11:01 PM   #61
diyAudio Member
 
Join Date: Apr 2009
Send a message via MSN to bacon665
Quote:
bacon665, interesting idea. I think if you use diodes, a 6AL5 is recommended rather than solid state, due to the 0.7V drop for commutation on most silicon diodes.
Thanks atma, That was a very quick drawing (less than 5 minutes) and i hadnt even begun to take into account componants other than the essentials to demonstrate the idea. Most people might shun me for this but something as simple as a FET operating in class A should easily compensate for the voltage drop and return the amp to class A correct?

I hadnt thought of vacuum diodes either, you rarely see them in amplifiers. in fact i rarely see them at all. Most of the ones i see are vertical and horizontal deflection tubes for CRT displays. Im going to post another tidbit i thought of after reading what me and YOU posted.

I know you formentioned that you don't like class B but to be honest I dont think i would mind it so much. I started to think what else amplifies bipolarly and then it dawned on me, one of the most common amplifier types in existence PUSH PULL! NPN's amplify positive voltage and PNP's amplify negative so why not take the amp and split it in two. You get the following circuit and im sure a myraid of problems out of my scope of knowledge to address.

Click the image to open in full size.

feedback anyone?
  Reply With Quote
Old 21st February 2010, 05:41 AM   #62
diyAudio Member
 
Miles Prower's Avatar
 
Join Date: May 2005
Location: USA
Blog Entries: 6
Quote:
Originally Posted by atmasphere View Post
FWIW there are some fundamental misconceptions here, the first being that all tubes are equal which they are not. From the very beginning of OTL technology in the early 50s, OTLs have used low-impedance tubes. That may sound oxymoronic to some, but we are talking about tubes that handle a lot of current and don't take much plate voltage to do it. Hence: the 6BQ6 above is a poor choice to make power, and the 6AS7G is an easy choice.
I have to disagree. The 6BQ6 was specifically designed as a TV horizontal deflection amp that can pull bigamps through the deflection coils at low Vpk's. These also have a max cathode current rating of 400mA, as opposed to the 125mA rating of the 6AS7. Run the 6BQ6 at 150Vdc on the plate and screen and you can hit 350mA of current in Class AB2. Not too bad, I'd say. Drive them with a source follower, and that would definitely work.

This type also sounds very good, even without any NFB applied. It's like a more powerful 6V6 in that regard. Of course, I'm not using 'em by the car load either, and unforch, they're not making any new ones.
__________________
There are no foxes in atheistholes
www.dolphin-hsl.com
  Reply With Quote
Old 22nd February 2010, 01:24 PM   #63
diyAudio Member
 
Join Date: Mar 2005
Location: Midwestern US
Quote:
Originally Posted by atmasphere View Post
...We implemented the first version of the voltage amplifier in 1985 so it predates Nelson by a few years. The CCS is a big deal- being 2-stage it allows the gain the 6SN7 sections to reject power supply noise and set up the current for the differential amplifier. Over a range of 100 to 128V with an unregulated supply the performance of the voltage amplifier is unaffected.

...the CCS won us so much extra performance that it was a no-brainer. BTW, by paralleling extra tube sections on the bottom you can decrease the plate load resistor to 1/2 or less of the value shown and get greater bandwidth, lower distortion and more gain all at once (usually you sacrifice one to get the other two). Of course the cathode resistor of the CCS would have to be adjusted to best gain/distortion at full whack.
Atmasphere,

I had read Walt Jung's 2007 results documenting the performance enhancements accruing to a cascoded current-source configuration and have enjoyed the sonic benefits of incorporating this CCS topology in a number of my own prototype designs. I thought I was being fairly progressive, yet I now find that this circuit enhancement has been a intrinsic element of the M60 OTL circuit topology for over 20 years!

I like the idea of cranking-up the standing current in the input stage, especially with the performance benefits that you've observed without incurring any degradation. Since the cascoded CCS is comprised of only seven components (one dual-triode, two resistors and four Zener diodes), it appears that the simplest/most-consistent approach to increasing the differential-pair idle current would be to just build a complete parallel cascoded CCS. If my calculations are correct, each cascoded CCS would supply 2.5ma of current, resulting in a combined 5ma (2.5ma per differential-pair device). The 6SN7 maximum plate current is 9-10ma, so the design would still be conservatively operating the tubes. Then we trim the input stage plate resistors from 100K-ohms down to 50K-ohms. It's so rare to get "something-for-nothing" design trade-off that I'll be very interested in evaluating this apparent "free-lunch" upgrade...

Doubling the input-stage idle current should also increase the ability to drive the driver-follower stage as well as increase the input-stage slew-rate, right?
__________________
Mullard EL34
  Reply With Quote
Old 22nd February 2010, 02:56 PM   #64
diyAudio Member
 
Join Date: Mar 2005
Location: Midwestern US
Quote:
Originally Posted by atmasphere View Post
I forgot...

...About the bias- 60 ma per section is class A2 with this circuit, due the grid current capabilities of the driver. With the plate voltage and this idle current, the tubes will not cut off without the amp also clipping.

You could run a little more current and less B+, as little as 100V...
Atmasphere,

I'll still working to get my head around some of the specifics of the M60's balanced-differential OTL output stage, so please bear with me on a few questions.

Current delivery will be key to powering the loudspeakers with an OTL design, so I've been scouring the 6AS7 datasheets (as well as a number of on-line discussions and books) to ascertain the tube's performance envelope with regards to current delivery. If I understand correctly, each half of a dual-triode 6AS7G is rated at maximum 125mA steady-state bias current, and can provide up to 300mA or so as a peak value. Therefore, the four 6AS7's per phase of the Class-A2 push-pull output (eight 6AS7's in total for both phases) would enable each phase to source/sink about 2.5A. That should be sufficient to deliver the 60-watts@8-ohm design specification.

While the configuration appears to conservatively operate the 6AS7's from a per-device current-load standpoint, it seems like the plate-dissipation is pushing things a bit (if my circuit analysis is correct... ). The 6AS7 is rated for a maximum plate dissipation of 13W/triode-section, but the 60mA dropped over the 150V across the plate would put the design at 9W of power-dissipation; it this calculation on my part correct? If so, is running the 6AS7 plates at 70% of capacity OK? Reducing the output stage to +/- 100VDC power-supply rails in each of the floating power-supplies would decrease my calculated 6AS7 plate dissipation to 6W (about 45% of maximum), which is more in line with my usual approach of running components at no more than 50% of their power ratings. Is this correct or is my understanding of the output stage faulty?

Class-A2 specifies that grid current flows for some part of the cycle, so am I correct in assuming that the 60mA idle-bias current per triode section is sufficient to ensure no switch-off of the device under normal operations, thereby precluding any cross-over distortion components?

According to some sources, the 6AS7G is designed to handle significant grid current, with the top part of what appears to be the anodes, between the top two mica plates, serving as a pair of radiators for the grids. Is this correct? If so, do all 6AS7 variants have this grid-current capability?

Again, thanks for your consideration of these questions from a DIY'er!
__________________
Mullard EL34
  Reply With Quote
Old 22nd February 2010, 04:53 PM   #65
diyAudio Member
 
atmasphere's Avatar
 
Join Date: May 2007
Hi Mullardel34,

The actual plate dissipation in this case is about 8.4 watts (140V on the plate, not 150V). The heatsink at the top of the tube is in fact the grid heatsink (on some tubes its at the bottom). The 6AS7A variant has a smaller heatsink and will quickly fail in this application!

You are correct- the 60mA value is sufficient to prevent cutoff. The M-60 is class A2 into 8 ohms but class AB2 into 4. OTOH if driving 16 ohms the output power goes to 80 watts and the output section runs cooler at the same time.

With regards to the CCS: you want to keep the impedance of that circuit high. We did explore paralleled sections. In a nut shell, they suck. The two stage circuit I showed supports 50K plate resistors, all you have to do is adjust the cathode resistance of V3 and the 6SN7 sections will easily support the extra current. We do this all the time in our MA-1- that is how it is set up.

Miles- The fact that you have to use so many tubes to make so little power suggests that the tube has a higher plate resistance, and unless you are doing something you did not mention, fairly low linearity. We've done a bit of work with the EL509, which is similar to the 6LF6, IMO if you want to use a sweep tube these are the ones to use. The 6LF6 is getting rather rare but the EL509 is still being made. There is a non-plate cap variant of it also, which is the tube that the KT-90 was developed from.
  Reply With Quote
Old 22nd February 2010, 10:02 PM   #66
diyAudio Member
 
Join Date: Mar 2005
Location: Midwestern US
Quote:
Originally Posted by atmasphere View Post
Hi Mullardel34,

The actual plate dissipation in this case is about 8.4 watts (140V on the plate, not 150V). The heatsink at the top of the tube is in fact the grid heatsink (on some tubes its at the bottom). The 6AS7A variant has a smaller heatsink and will quickly fail in this application!
Atmasphere, thanks for confirming that (1) I had correctly determined the means of calculating the plate dissipation (even if I did use the wrong voltage-drop value across the plate... ) and (2) that we do need to be cautious when selecting specific 6AS7 variants so that the tubes we use have sufficient grid-current heat dissipation capacity. If we use power-supply transformers that result in slightly different power-supply rail voltages (there aren't many readily-available power transformers with 106VAC secondaries...), we'll need to adjust the resistor values in the driver-stage bias network accordingly in order to maintain the fixed-bias value necessary for the 60mA design target.

Quote:
You are correct- the 60mA value is sufficient to prevent cutoff. The M-60 is class A2 into 8 ohms but class AB2 into 4. OTOH if driving 16 ohms the output power goes to 80 watts and the output section runs cooler at the same time.
Excellent! It makes perfect sense that a fixed bias level that configures Class-A2 operations into 8-ohm loads would be insufficient to keep the output devices in conduction when driving a 4-ohm load. I assume that, when presented with a 4-ohm load, that would be a good situation in which to use the Speltz Zero-autoformers to elevate the effective loudspeaker impedance presented to the OTL output to something that would preserve the Class-A2 mode of operation?

Quote:
With regards to the CCS: you want to keep the impedance of that circuit high. We did explore paralleled sections. In a nut shell, they suck. The two stage circuit I showed supports 50K plate resistors, all you have to do is adjust the cathode resistance of V3 and the 6SN7 sections will easily support the extra current. We do this all the time in our MA-1- that is how it is set up.
The failure of paralleled tubes is interesting, given the successful use of the "composite triode" configuration in designs such as the conrad-johnson ART product series. Any thoughts on why the paralleled tube experiments were incompatible with the (obviously) successful M60 input-stage design?

OK, so we just drop the 37.5K-ohm resistor value in the cathode of the current-source (v3) to 18.75K-ohms (bringing the total input-stage current to 5mA), then scale-down the cascode (V2) plate resistors to 50K-ohms. If it's good enough for the MA-1, it should be just fine for a hot-rodded M60; all input-stage device current levels and heat-dissipation capabilities are still being safely and conservatively operated. Literally, "cool", man!

I think that answers my questions on the signal-path topology of the M60 OTL. I'm still just floored by the elegant sophistication of the design, harvesting massive performance from a minimum of circuit elements!

I'll go perform a bit of research and make an initial high-level pass at developing a power-supply for the OP and the others (including me...) who are interested in building a non-commercial implementation (we've gotta respect the intellectual property rights of those willing to share knowledge with the DIYaudio community... ) from off-the-shelf, commercially-available components.
__________________
Mullard EL34
  Reply With Quote
Old 22nd February 2010, 10:42 PM   #67
diyAudio Member
 
Join Date: Apr 2009
Send a message via MSN to bacon665
any thoughts on the effect of a floating smps might be on the M-60? seems to be the most logical way of getting that 160v other than going 120:220 step up and then using a regulator stage to drop and regulate it to 160
  Reply With Quote
Old 22nd February 2010, 11:18 PM   #68
diyAudio Member
 
Join Date: Mar 2005
Location: Midwestern US
Quote:
Originally Posted by bacon665 View Post
any thoughts on the effect of a floating smps might be on the M-60? seems to be the most logical way of getting that 160v other than going 120:220 step up and then using a regulator stage to drop and regulate it to 160
I'll let others address the viability of floating SMPS power-supplies for this project; personally, I prefer conventional linear power-supplies in my audio equipment...

As I mentioned in my previous post, I'm making an initial effort to research the "hard" items necessary for a suitable power-supply to support a DIY Atma-Sphere M-60 OTL monoblock or stereo power-amplifier. I've already unearthed a number of off-the-shelf power transformers which can provide the necessary secondary voltage/current capacities. I'll be presenting these items in a subsequent post; however, I've gotta address the requirements of my "day job" first...
__________________
Mullard EL34
  Reply With Quote
Old 23rd February 2010, 02:38 AM   #69
diyAudio Member
 
atmasphere's Avatar
 
Join Date: May 2007
Mullardel34, The Speltz ZERO is based on a product we previously made called the Z-Music transformer (although it was really an autoformer). That was precisely its function- to allow the amp to drive lower impedance loads that it would otherwise be heavily compromised in doing. So the autoformer only has taps for low impedances, since the amp can drive 8 ohms directly. IOW, its a problem solver.

Usually in a CCS you want to keep your impedances as high as possible. Its the same idea as if you had no CCS: high resistance resistor and super-high B- voltage (like maybe -1000V). That is what the CCS is doing for you but it can degrade the performance if you reduce its impedance. The main thing is make sure that the tube sections are operating in their linear range. Some years back a few people were trying out 6SL7s in place of the 6SN7, which might have worked were it not for the high currents. So what was happening was that the grid of the tube was positive w/respect to the cathode... sure, it sounded different, but it was not **better**.

I've often thought a SMPS would be fun to play with, but you would have to build something up custom. Modern SMPS have got much better noise figures so I think you could pull it off. Wouldn't have happened 20 years ago though. The thing to keep in mind is that the output section has very wide bandwidth- one time I used one as a linear amplifier boot for a CB radio It worked great and didn't have a lot of harmonics. IOW you really don't want that bandwidth interacting with HF/RF noise in the power supply, unless you like a nice tan
  Reply With Quote
Old 23rd February 2010, 07:00 AM   #70
markusA is offline markusA  Sweden
diyAudio Member
 
markusA's Avatar
 
Join Date: Dec 2009
Location: Gothenburg
Blog Entries: 1
Pardon me making a maybe not so gifted question, but is it possible to insert a volume pot (let's say 10kohm ldr or something along those lines) somewhere without losing the high input impedance? I'm thinking remote controlled pots, one on each monoblock. Just toying with different ideas at this stage. Getting rid of the preamp would be a nice thing and I only use one source anyway.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
New to Tubes, just bought first tube amp josefgabriel Tubes / Valves 28 7th April 2010 07:20 AM
FS: 1626 Tube Amp, Van Alstine pre, tubes, crossover. xecluded Swap Meet 1 30th April 2009 07:27 AM
Tube Noob! need help with tubes lots of tubes therapy_fan Tubes / Valves 15 29th January 2009 09:01 PM
Adding more power tubes and make a tube amp switching them? Nicko_Lps Tubes / Valves 3 7th January 2009 12:31 AM
"The Fisher" tube amp 7189 output tubes laflamme Tubes / Valves 4 28th August 2004 12:40 AM


New To Site? Need Help?

All times are GMT. The time now is 06:35 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2