12AX7 to 6SN7 conversion: Any thoughts, comments, concerns, advice, etc?

[Wharfcreek]

I've been building Mr. Gillespie's little Magnavox 8600 revision amp numerous times......and consider it both a GREAT amp when done, as well as just a fun project! So nice to see something like this come around and be such a great 'learning tool' not to mention providing such satisfaction when completed. Again, thanks to Dave for sharing his work on this revision.

So, now it comes to some of the things I've done to actually 'modify' Mr. G's modifications....which up to this point has been pretty minimal. I've worked pretty much exclusively with 5K PI OTs with 8 ohm secondaries, or combination 4/8 secondaries....but I've maintained the circuit as Dave developed it with using the feedback off an 8 ohm output. But, I've been asked to now do two things in building one of these amps: 1) change from the 8 ohm secondary to a 16 ohm secondary. And 2) substitute a 6SN7 for the 12AX7. While the feedback issue will be a necessary bridge to cross, the main subject of this post is to see what kind of thoughts, suggestions, comments might come from attempting to make this change in the circuit? For the record, I have checked the filament current draw issue out, and that doesn't seem to be a problem as I'm using all new parts for this build, including a new power transformer, and it's well capable of delivering all the needed 6 volts power to the filaments of the 6SN7 as well as the EL84s. I could actually substitute some 6V6 tubes for these as well, and make it an 'all octal' amp......and I might do next. But, for now, I have to cross the bridge of changing the 12AX7 to a 6SN7, and I guess I just wonder what kind of 'trouble' I could be making for myself here? As noted in the circuit, the plate voltage as designed is 137 VDC.....which is well within the limits of the 6SN7. But, is it enough? Should that get 'increased' some? If so, it seems the 470K resistor ahead of the plate could be 'adjusted' to accommodate that. Likewsie if the 6SN7 would actually perform better at an even lower voltage, again, the 470K resistor could be increased as well.

Then there's the issue of the cathode resistor; should it remain at the designed 820 ohms, or while that may be great for the 6BQ5, the 6SN7 would perform better with a different value in this position? I'm hoping that someone with some familiarity with this tube will have some advice for me. But, I also am wondering if I've forgotten to take anything else into consideration? Like that say, some times it's what you 'don't' ask that causes you the problems.

So, that's it! If anyone has any comments, please post 'em!! And if you want to make any suggestions regarding any adjustments to the feedback values (currently 10K and 300pf)....that would be great as well.

Thanks All!!

Tom D.

 

[century tek]

Just kind of skimmed your post, but wouldn't you want to use a 6SL7? The 6SN7 is more like a 12AU7 but you probably already knew that.

 

[Wharfcreek]

100-tek: Actually, I had not checked that aspect of things......as the 6SN7 was the 'requested' tube. But, now that you've mentioned it, I believe I did come across this once in the past and was told the same thing; that the 6SL7 was basically a direct 'octal' sub for a 12AX7. In fact, if I recall correctly, I believe I even read an article some place where the 'history' of tubes was being discussed, and the development of the 'miniature' tubes (ie the 9-pin and 7-pin types) were done as a cost-saving and 'space'-saving' effort. I think that article also mentioned that it was actually the reverse of what I just implied previously....... it was more that the 12AX was the 9-pin 'sub' for the octal 6SL7. The 12AX being the 'mini' replacement version of the 6SL. So.....maybe if I use that tube, all the electrical values on plate and cathode can remain the same?

Thanks for that comment........very helpful!

TSD

 

[knockbill]

+1,,, I've used 6SL7s in quite a few builds that called for 12AX7/7025 with good results... As CT mentioned, 6SN7 is closer to 12AU7.... If you want to use octals, 6SL7 should work...

 

[FlaCharlie]

I believe I did come across this once in the past and was told the same thing; that the 6SL7 was basically a direct 'octal' sub for a 12AX7. . . . The 12AX being the 'mini' replacement version of the 6SL. So.....maybe if I use that tube, all the electrical values on plate and cathode can remain the same?

In terms of gain (mu) the octal 6SL7 is actually more similar to the 9 pin 5751, they both have a mu of 70 while the 12AX7 is 100.

The 6SN7 has a mu of 20. There are a few 9 pin miniatures with the same gain: 12AU7, 12BH7, 6CG7. Some of the other parameters are likely different to some degree so check the data sheets.

I assume you're building this in a different chassis but if you're doing that to accommodate the use of an octal socket you could just use one of the 9 pins instead.

I recently modded the similar 83xx series Maggie to use 12BH7s as outputs, which allows the use of the stock OTs. I'm amazed at how nice it sounds.

http://www.audiokarma.org/forums/in...0-or-8600-conversion-to-12bh7-outputs.824135/

I left the 6EU7 in place but I may go back in at some point and rewire the socket for a 12AX7, which would allow the use of 5751, any of the 12xx7 series, and even using another 12BH7 as the input tube.

 

[primosounds]

Wharf you should consider building an all octal amp. This would be visually much more appealing than the to have 9 pin power tube and octal input tube. Our dear departed friend Keggar had built a 6l6 based amp with 6sl7 input tube. I have lost the address for the schematic but his design/schematic was actually based on this one with small differences.
http://rh-amps.blogspot.com/2013/01/rh-universal.html
This amp is an excellent sounding performer and is worthy of the finest components.

 

[gadget73]

yeah, the 6sn7 is a very different tube than a 12ax7. You can make one sub for a 12au7 without much fuss, but a 6SL7 is the better octal sub for a 12AX7. I suspect you won't get enough gain out of the thing with a 6SN7 to work correctly. The 6SN7 isn't exactly an octal version of the 12ax7, it does have less gain, but its a whole lot closer. Its not too far off a 5751 though.

If you're just looking for something different, try a 6AV6. Its half a 12AX7, so functionally no difference worth talking about but you get a cheap as chips 7 pin tube per channel. It does have a pair of diode plates you'll need to tie to the cathode, but thats just a jumper wire on the socket.

Feedback values are fairly easy to convert. Resistor size you divide by 0.7 to switch from 8 to 16. If its got a cap across the resistor (don't remember), multiply that by 0.7. The resistor gets bigger, the cap gets smaller. Still need to confirm proper stability and HF response since it won't be the same transformers it was originally designed for but that ought to get you in the ballpark.

 

[Wharfcreek]

Actually, my plan was to incorporate the 'octal' base driver tube into the build as per the request of the individual who has asked for it this way. BUT.....in doing this, my 'plan'....was to do yet another on of these DG-SE1 builds using my 5AR4 rectifier PS circuit (that I've now used about 15 to 20 times), the Octal driver, and 6V6 output tubes. I think that's a pretty easy 'option' in converting from 9-pin tubes to an octal build. But, I'm under the impression that something like a 6L6 or possibly a KT88 or 6550....those tubes all require a bit more 'drive' than what one stage of a 12AX7 (or 6SL7?) is able to deliver On the other hand, I'm told that something like a 7591 is equally 'easy' to drive as the 6V6 or EL84/6BQ5 tube. Honestly, I'd LOVE to build something with a bit more output power.....and perhaps a move to a 7591 based SE amp might be a good starting point? But.....yes!.....and 'all octal' build is going to happen in one form or another. I just think I need to do this one step at a time.....starting with 'subbing' the octal for the 9-pin driver.....or the 6SL for the 12AX. Once I get that far, then it's on to the output section. BTW, correct me if I'm wrong, but a 6L6 build would also involve a pretty major power supply revision as well, would it not? Aren't we talking a plate voltage level at more like 350 to 400 volts vs the 6V6 at the 250 seen in Dave's revision of the 8600 amp?

Tom

 

[Wharfcreek]

Thain, thanks for the comments. Just as an aside to all this, I have 'subbed' the 12AX in the Dave's circuit with a variety of 'other' tubes, including AT, AU AY, and AZ. What's interesting is that in making these substitutions as just a way to play around with the amp some, I found that the inter-action between amp and 'pre-amp' or driver source had some almost 'favorable' results when using different tubes. I'd be remiss if I implied that the amp was actually working 'better'.....and perhaps some appropriate testing would reveal some deficiencies that have a clear 'adverse' affect. BUT, that said, it did seem to 'work' better in some circumstances. I'd guess this might have been a function of load impedance levels, pre-amp output levels, etc. But.... this is clearly something I have no real ability to validate one way or the other. And, the only reason I'm even considering the 6SL7 now is because it was asked for. And, I do think it would be interesting to have an 'all octal' version of the DG-SE1 to compare to a 9-pin version. My guess is that they may be indistinguishable. I hope to find out.

 

[FlaCharlie]

I do think it would be interesting to have an 'all octal' version of the DG-SE1 to compare to a 9-pin version. My guess is that they may be indistinguishable. I hope to find out.

Dave's mod is designed to be an improvement on a particular amp. I suspect if he were building a SE 6V6 from scratch that he might make some different parts choices, especially the OTs, since they wouldn't need to be small enough to be mounted on the tiny chassis of the 8600 Maggie. A lot of people like the 6V6 in triode so you might consider that if you don't need as much power.

You could also build a loctal amp using 7C5s (6V6 equivalent) and a pair of 7B4s on the inputs. The 7B4 has a mu of 100, just like the 12AX7 but it's a single triode so you'd need two. NOS Loctals are relatively cheap - total cost of tubes would be $20 to $24, not counting the rectifier. I'd probably go for a 6BY5 there. Cheap and slow start like a 5AR4.

 

[Wharfcreek]

FlaC, I think you're probably right about how DSG would go about a complete 'scratch' build around 6V6 tubes. But, the interesting thing to me is that the 6V6 and 6BQ5 are so frequently interchanged in different designs. I think you can buy 'adapters' to go either way from one socket to the other tube. I know Dave was working with the 8600 as more of a project for he and his Grandson, vs designing an amp from scratch. However, with all the optimization done within his revision, I see the circuit as an easy 'scratch build' vs modifying an 'original' 8600 chassis. I've done a couple of dozen builds of this circuit now, and I think only 2 have actually been on the Magnavox chassis. All the others have been on something else; either some other re-purposed chassis, or starting out 'all new' with Hammond chassis or similar. I've done most with 5 volt rectifiers, mostly 5AR4s, but I've yet to change from the 12AX7 or 6BQ5. So, I want to see what, if any, differences I 'hear' in the amp when using something like the 6SL7....but specifically an 'octal' driver tube. Then I want to do a 6V6 version....and if the 6SL7 works out, it will be an 'all-octal' amp. BTW, I've been pretty much doing all my late builds of this circuit using the Edcor GXSE OTs, and have been quite satisfied with the results of those. For the price, they're a pretty good deal! I just think an 'all octal' version would be a unique approach to an otherwise 'GREAT' design effort by Dave.

 

[thorpej]

The 6SN7 has a mu of 20. There are a few 9 pin miniatures with the same gain: 12AU7, 12BH7, 6CG7. Some of the other parameters are likely different to some degree so check the data sheets.

FWIW, 6CG7 / 6FQ7 are electrically equivalent to 6SN7. And the only difference between 6CG7 and 6FQ7 is an internal shield between the two triodes that's brought out to a pin for grounding in the 6CG7.

 

[Wharfcreek]

So it seems the 6SL7 is the best 'octal' to consider when attempting to use an octal to replace a 12AX7.

 

[primosounds]

..... BTW, correct me if I'm wrong, but a 6L6 build would also involve a pretty major power supply revision as well, would it not? Aren't we talking a plate voltage level at more like 350 to 400 volts vs the 6V6 at the 250 seen in Dave's revision of the 8600 amp?

Tom

Well it kinda depends. If you check the tube specs for the different tubes, various operating points can be chosen with different plate and g2 voltages in single ended use. The power supply for 6l6 requires higher current for the B+ and slightly more heater current. For EL34 and 6550 type tubes the heater and B+ currents again need to be increased. Actual drive requirements for the driver tube is still within the capability of the 6sl7 or 6sn7 types, so the driver circuit simply needs to be tweaked to have the correct characteristics. Seriously, check the rh universal amp. I have built many of this type and i find the sound quality to be exceeded only by much more expensive SET 845 or 211 amps.

 

[Wharfcreek]

Primo, I'm familiar with Alex's work......and I know he has some excellent designs, ideas, and results. However, at the moment I'm not really looking for alternative designs or projects.. I can see doing something like the RH amp at some point, but I want to stress that I'm not just looking for a project to do here.....I'm specifically looking to try a different set of tubes in Dave's 8600 revision with optimum results....then compare that to the original revision build. If the question to that is 'Why?'...the answer is simple: because I've been asked to by someone who wants an amp built that way.....and I have a similar interest for myself. I'm pretty confident that I can just change out the sockets and pin wiring and replace the EL84s in the circuit with 6V6 tubes and have the output section done. But, changing the 12AX to a 6SL7 will be a new one for me. I'm not all that familiar with that tube so I don't know if the 137 plate volts that Dave shows, or the 820 ohm cathode resistor will 'operate' that tube in a suitable manner in this circuit.

 

[gadget73]

Once you change the tube lineup, you've basically re-invented the wheel. The bias points on the tubes are going to be different and things will need tweaking to make it happy. The basic topology really won't change just because there are only so many ways to actually build a working amplifier without adding in a lot of stuff just for the sake of complexity or to meet some very specific need.

 

[Wharfcreek]

Not to muddy my own water here, but in doing some more checking I've come across the 6SC7.....which appears to be pretty close to the 12AX characteristics as well. Between it and the 6SL7, I'm wondering if perhaps it's not a slightly better choice? Comments?

 

[Wharfcreek]

Gadget, I think your comment rather gets to the point of my quest here. In swapping the 6BQ5s to 6V6 tubes, there's really not much of a change needed to 'adjust' the circuit for the change in tubes, or at least that been my experience with those two tubes so far. Most of this experience comes from Guitar Amp builds.....but I've done it in 'home audio' units as well. Kinda like swapping 7868s for 7591s. I was just hoping that I could do the same with some kind of octal base driver tube such that the two basic operating points (plate voltage and cathode resistor points) could be kept at basically the same values that are shown on Dave's schematic for his 8600 revision. That said...... I think the same concerns would apply here if swapping the 12AX for a 12AT or AU, etc. Yet, I've simply plugged those in and, as I think I indicated above, had some interesting and 'not bad' results. Again, the circuit as Dave has developed isn't designed for the AT or AU, but it does work with them......and in some respects I actually found the results of the change to be favorable in some ways. However, I've always returned to the AX simply because it's designed for it and I KNOW it's working properly with it. Anyway...... if changing the cathode resistor to something other than the 820 ohm unit, or if raising or lowering the plate voltage is advisable for use with something like the 6SL7 or 6SC7....then that's 'doable' as well. I just need to know which way to go....

 

[knockbill]

I built a phono amp (GE designed, IIRC, Fisher also used it, I believe),,,it uses 6SC7s.... It was used with an early GE cartridge that didn't have RIAA EQ... Got help designing the RIAA section, and it turned out well...
6SC7 has common cathodes, if that matters for your new build... Also,,, 6V6 is a much closer match to 6AQ5,, than 6BQ5,,,
I've found some 6SC7s to be microphonic and/or hum... Quiet ones sound very good tho...Make sure you ground pin1,,, Its the metal can/shield...
Seems changing both preamp and output tubes will get you a completely different design,,,, But what do I know!!!!

Have fun...

 

[FlaCharlie]

But, changing the 12AX to a 6SL7 will be a new one for me. I'm not all that familiar with that tube so I don't know if the 137 plate volts that Dave shows, or the 820 ohm cathode resistor will 'operate' that tube in a suitable manner in this circuit.

In swapping the 6BQ5s to 6V6 tubes, there's really not much of a change needed to 'adjust' the circuit for the change in tubes, or at least that been my experience with those two tubes so far. Kinda like swapping 7868s for 7591s. I was just hoping that I could do the same with some kind of octal base driver tube such that the two basic operating points (plate voltage and cathode resistor points) could be kept at basically the same values that are shown on Dave's schematic for his 8600 revision. That said...... I think the same concerns would apply here if swapping the 12AX for a 12AT or AU, etc. Yet, I've simply plugged those in and, as I think I indicated above, had some interesting and 'not bad' results. Again, the circuit as Dave has developed isn't designed for the AT or AU, but it does work with them......and in some respects I actually found the results of the change to be favorable in some ways. However, I've always returned to the AX simply because it's designed for it and I KNOW it's working properly with it. Anyway...... if changing the cathode resistor to something other than the 820 ohm unit, or if raising or lowering the plate voltage is advisable for use with something like the 6SL7 or 6SC7....then that's 'doable' as well. I just need to know which way to go....

As Knockbill mentioned, the 6V6 is closer to the 6AQ5 than the 6BQ5 while the 7868 and 7591 are the same tube with different bases. And, yes you can pop in various 12A*7 tubes in place of a 12AX7 because the pinouts are the same but that does not mean they are optimized for the circuit. Every tube has its own characteristics which is why you need to refer to the data sheet and use the plate curves to find a suitable, optimal, operating point. The operating point for the 6SL7 will be influenced by the operating point you use for the 6V6. Here are the data sheets:

https://tubedata.altanatubes.com.br/sheets/093/6/6SL7GT.pdf

https://tubedata.altanatubes.com.br/sheets/093/6/6V6GT.pdf