Spreadspectrum and kenpeter,
THANKS for the input, so it's pentode PP.
Don't you guys think that triode connecting the output tubes by attaching screen to plate, and connecting an 8 ohm loudspeaker to the 4 ohm tap (primary impedance is 4k6 then) would yield a better sounding amp? Could easily be done with this type of amp. Then we have a 35 watts triode PP amp. A bit more current so a bit more of class A, a bit less GNFB....
THANKS for the input, so it's pentode PP.
Don't you guys think that triode connecting the output tubes by attaching screen to plate, and connecting an 8 ohm loudspeaker to the 4 ohm tap (primary impedance is 4k6 then) would yield a better sounding amp? Could easily be done with this type of amp. Then we have a 35 watts triode PP amp. A bit more current so a bit more of class A, a bit less GNFB....
It appears that Jan V. has all bases covered!
Circlotron:
Circlotron Monoblock Power Amplifier
Twin Coupled:
KT66 Class A CFB amplifier
Mac:
BigMac, Unity coupled output amplifier
Ever get that feeling of being a small boy again in the model train store/toy shop!
-----------------
"Don't you guys think that triode connecting the output tubes by attaching screen to plate, and connecting an 8 ohm loudspeaker to the 4 ohm tap (primary impedance is 4k6 then) would yield a better sounding amp?"
Well I guess you could do that. The 50 % CFB would make that kind of redundant, since it effectively turns the pentode mode into a Mu=2 triode already.
Circlotron:
Circlotron Monoblock Power Amplifier
Twin Coupled:
KT66 Class A CFB amplifier
Mac:
BigMac, Unity coupled output amplifier
Ever get that feeling of being a small boy again in the model train store/toy shop!
-----------------
"Don't you guys think that triode connecting the output tubes by attaching screen to plate, and connecting an 8 ohm loudspeaker to the 4 ohm tap (primary impedance is 4k6 then) would yield a better sounding amp?"
Well I guess you could do that. The 50 % CFB would make that kind of redundant, since it effectively turns the pentode mode into a Mu=2 triode already.
Last edited:
Yes, the best sounding tube amps I have heard have been the constipated kind. Like Macs. And although tubes typically drive speakers with plenty of authority, distortion, or at least clipping, are bad things. Good things in a Marshall, bad things in HiFi. Tube HiFi should be idling, always.
Did we answer the sheer power question raised today? The 275 made 150 watts of authority. It's a brute. I have four KT-88's making sixty. I believe you guys when you say there are superior modern designs and parts, and I'm not intending to build a Mac. But you know.....they did made smooth, wonderful-sounding, brutish amps.
The Mac can do the big Watts because its rather efficient using the class B biasing.
Another way to get good efficiency is by using screen grid drive. SY can comment on that, he's got one. George (Tubelab) has used that too, like maybe 150 Watts out of a couple of 24 Watt tubes! Have to be into Nuclear Cooling I think. But, more reasonable and still impressive Watts are practical. I think George will probably come up with something interesting using Screen grid drive (maybe G2/G1 drive) and Schade feedbacks this Winter. I'm going to look at Schaded Elliptron (halfway between Circlotron and Twin-C) this Winter, using the new "magic" feedbacks and G2/G1 drive. Orthogonal feedbacks are just too interesting now.
Elliptron scheme:
http://www.diyaudio.com/forums/tubes-valves/19624-elliptron-schematic.html
OOPs, correction: Looks like the Twin Coupled link I posted above to Jan's site is really another Mac type, its all on one output transformer, instead of two. But that would mean that Sowter already makes a 60 Watt Mac type OT. There's a possibel start. Oh, wait, it's only 18 Watt (class AB rated).
Thought I saw 60 Watt someplace there.
Another way to get good efficiency is by using screen grid drive. SY can comment on that, he's got one. George (Tubelab) has used that too, like maybe 150 Watts out of a couple of 24 Watt tubes! Have to be into Nuclear Cooling I think. But, more reasonable and still impressive Watts are practical. I think George will probably come up with something interesting using Screen grid drive (maybe G2/G1 drive) and Schade feedbacks this Winter. I'm going to look at Schaded Elliptron (halfway between Circlotron and Twin-C) this Winter, using the new "magic" feedbacks and G2/G1 drive. Orthogonal feedbacks are just too interesting now.
Elliptron scheme:
http://www.diyaudio.com/forums/tubes-valves/19624-elliptron-schematic.html
OOPs, correction: Looks like the Twin Coupled link I posted above to Jan's site is really another Mac type, its all on one output transformer, instead of two. But that would mean that Sowter already makes a 60 Watt Mac type OT. There's a possibel start. Oh, wait, it's only 18 Watt (class AB rated).
Thought I saw 60 Watt someplace there.
Last edited:
I beg to differ... GNF may play a part, but it is mostly due to the plate resistance of the tubes matching up to the output transformer / speaker.
I have yet to see a pentode amp reproduce bass notes as well as a triode amp... Feedback or no Feedback.
MAC amps are configured as pentode. As Smoking-Amp stated, MAC did a neat trick with their output transformer connecting the screens in such a way to keep the screen voltage constant in relation to the same tube's cathode. This is the ideal operating mode for a pentode.
1000 ohms is the primary impedance for each of the windings for the MC30, and MC40. This effectively gives the 6L6's in these amps a working load of 4000 ohms. The MC60 and MC75 (old) have 600 ohm primary output transformers for an effective load of 2400 ohms. These loads are not too low for 6L6, 6550, KT88 in class "B" mode (actually class "AB2") that the MAc operates in.
All pentode amps are a mismatch to the output transformer. 6L6's have a plate resistance of 30,000 ohms give or take. 6550, KT88's have a plate resistance of 15,000 ohms. The primary impedance of an output transformer would have to be around 60K to 120K ohms for a 6L6 and 30K to 60K for the other 2 tubes mentioned to get good damping. Gross mismatch indeed! This is why a pentode amp will never outperform a triode amp and why pentode amps can never get a handle on driving the speakers like a triode amp can.
Plate resistances of your common audio triodes for the most part are less than 1000 ohms. 4 X this is 4000 ohms; an easy impedance to make for an output transformer.
Daniel
This defeats the whole purpose of the MAC design. F. McIntosh and G. Gow set out to make a 50 watt amp with only 2 output tubes and have distortion figures that were unmeasurable (at that time). What they did was genius. They didn't come out with this topology to be tricky; they did it because it met their goals. The many gain stages were needed. All the feedback was needed. The tricked out output topology was needed to meet their goals.
This was in the late 40's. The only other serious High Fi amp at that time was the Williamson, which was good for approx 10-15 watts. Brook had the 12A3 (?) that put out about 10 watts and it was a true triode amp using 2A3's.
Did anybody see this link at the bottom of the Big Mac link? A great discussion of the MAC circuit by the inventors themselves:
http://www.veiset.net/tech/Mac1949.pdf
Daniel
http://www.veiset.net/tech/Mac1949.pdf
Daniel
These pentodes are Schaded by cathode feedback into triodes with an
effective Mu of 2. When you do that, plate resistance drops exactly like
a low Mu triode. Much lower than if you had strapped plate to screen.
Rp=Mu(effective)/gm
Mu(eff) = gm1/(CFB*gm1+UL*gm2+gmp)
Then global Fdbk lowers it even more.
--------------------------------------------
Some questions for Rosinante:
How many output Watts are you looking for here?
And do the tubes have to be standard Audio ones like KT-88 ... or are NOS Sweep tubes acceptable? Reason I ask is that the usual audio tubes use high screen grid voltages and so are not useable with screen drive.
And are parallel output tubes acceptable? Or this has to be one set of big ones? 36LW6 Sweep say, 40 Watt diss, 200+ Watt output per pair (George!).
Then I guess there are the big Xmitting type tubes, but the B+ voltages get way too dangerous then in my opinion. Not to mention the difficulty getting a custom HV OT made as well.
Oh, and then does this need to be high feedback/low-low distortion. Reason on this is high feedback will demand a lot of test equipment to get stabilized, and the cheapo OTs will not hack the stability issue at all.
Well, these should all be questions for a later amplifier really. I would still recommend doing something easy the first time, likely to work, and to sound good by most opinions. Then you will have a better feel for what features are important, try some upgrades, tweeks, know where it's worth to spend $$$$ on the next big one.
Another issue would be the power supply. I guess most people just buy the appropriate power components for the amplifier at hand with the idea that the whole thing could be sold later if something new is developed. The other approach, which I use, is to just build a big power supply with a Variac on the primary to adjust B+ for the amp at hand, then that will work for any future amp. You only run two amps at a time anyway. I hate building power supplies. This probably makes more sense if you plan on endlessly building new "idea" amps. Another way is to just get some nice variable bench supplies. An off the shelf solution.
Mu(eff) = gm1/(CFB*gm1+UL*gm2+gmp)
Then global Fdbk lowers it even more.
--------------------------------------------
Some questions for Rosinante:
How many output Watts are you looking for here?
And do the tubes have to be standard Audio ones like KT-88 ... or are NOS Sweep tubes acceptable? Reason I ask is that the usual audio tubes use high screen grid voltages and so are not useable with screen drive.
And are parallel output tubes acceptable? Or this has to be one set of big ones? 36LW6 Sweep say, 40 Watt diss, 200+ Watt output per pair (George!).
Then I guess there are the big Xmitting type tubes, but the B+ voltages get way too dangerous then in my opinion. Not to mention the difficulty getting a custom HV OT made as well.
Oh, and then does this need to be high feedback/low-low distortion. Reason on this is high feedback will demand a lot of test equipment to get stabilized, and the cheapo OTs will not hack the stability issue at all.
Well, these should all be questions for a later amplifier really. I would still recommend doing something easy the first time, likely to work, and to sound good by most opinions. Then you will have a better feel for what features are important, try some upgrades, tweeks, know where it's worth to spend $$$$ on the next big one.
Another issue would be the power supply. I guess most people just buy the appropriate power components for the amplifier at hand with the idea that the whole thing could be sold later if something new is developed. The other approach, which I use, is to just build a big power supply with a Variac on the primary to adjust B+ for the amp at hand, then that will work for any future amp. You only run two amps at a time anyway. I hate building power supplies. This probably makes more sense if you plan on endlessly building new "idea" amps. Another way is to just get some nice variable bench supplies. An off the shelf solution.
Last edited:
Oh, another item is plate caps on the output tubes, for appearance, and safety issues. Most of the standard Audio tubes are capless. The Sweep tubes are available up to 24 or 28 Watt plate diss. without caps. Caps though up to the 40 Watt Sweeps, but are generally available as insulated caps. Some builders like the high power look of the caps I think. Oh, and sweep tube watt diss. ratings can usually be upped by 20% or so when comparing to audio tube diss. ratings. Xmitting tubes are almost always capped, and often require dangerous uninsulated finned cooling caps. (You need a cage over the tube section then for sure.)
Last edited:
Can you please post layout/winding schematic, or at least describe it? Thanks !
The data measurements do look very nice indeed. 26 dB of feedback would be expected to fix most of the crossover glitching, but not totally. The complete lack of a residual crossover glitch in the analyzer trace tells me it's not running in class B though. They finally wised up and biased it up to optimum class AB for audio. No where in the article could I find any info on the idle current of the outputs.
Classic Mac amps run class AB2; I don't know why people keep talking about MAC amps and class "B" because they run in class "AB2".
In order to make the amps user friendly MAC eliminated the bias adjustment pot. It's just a fixed resistor. My MC-60 clones run anywhere from 30ma to 50ma on each output tube, depending on what brand they are and how they bias up. All voltages are correct in this amp according to Mac's schematics. The MC75 / 275 is probably somewhere near that value. IIRC the MC30's I had biased at around 30ma for each output tube. Now take a look at the voltage going through the 12AX7 driver; over 450 volts! That was not a place where McIntosh was conservative.
I am reluctant to post a complete winding schematic as the amplifier is still a commercially available unit.
What struck me however was that the winding was not "tidy".
You can start winding a transformer bifilarly, but when no interleaving is applied the windings start to take their own route after a couple of layers which was the case here; the outer layers on the coil could actually not called bifilar anymore.
Another thing that struck me was that there was also no interleaving/isolation between primary and secondary sections. The only isolation between the high voltage primary windings and the secondary (loudspeaker) windings was the wire isolation itself, and though the isolation can withstand these voltage potentials I don't think it is good practice. The "untidy" winding might be the main cause of the rather high failure rates of these output transformers because there is room for the windings to resonate which damages the wire isolation.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.