Now what do you mean by a 1/2 winding!!
The winding is either a full cathode winding or a full plate winding. The tube drives both in series. The other bifilar wire in each pair is used by the other tube. In each bifilar pair, one wire is a cathode winding, and the other wire is a plate winding (but for the other tube).
The winding is either a full cathode winding or a full plate winding. The tube drives both in series. The other bifilar wire in each pair is used by the other tube. In each bifilar pair, one wire is a cathode winding, and the other wire is a plate winding (but for the other tube).
Last edited:
Its clear you wasting everybody's time fishing for information on a transformer you are speculating about..
1/2 winding because Pieter made his voltage measurements with respect to GROUND... and GROUND is at the 1/2 of both PLATE and CATHODE winding Center-Taps...
Once again.. If the windings are in SERIES, as you claim, then are friend Pieter from Netherlands can simply measure the AC voltage of these series windings and confirm your theory...since he has the test set-up already....
I bet it can't be measured because it is not there to measure....proving the windings are in PARALLEL...
Or do you trust me to make a VIDEO to post to YouTube showing all measurement voltages with a diff probe in real-time ????
You define the test points and I will measure ....Is that fair ??
Last edited:
I really don't care. The Mac OT is trivial in concept. Just look at post 58 to see how it works. If they don't have it wound that way, you can just throw the things out, worthless junk, they didn't get it right.
The Circlotron runs circles around it. So much easier to wind than the Mac OT besides.
But lets have Pieter make the measurement.
But as I mentioned, we already have the data to decide. The cathode AC voltage and the plate AC voltage. They either sum to twice for one tube, or to zero. Zero means zero gain, no output at all. So they must sum to twice.
If you are getting a gain of 0.75, it just means you need new tubes.
The reason no one can wind one is they don't have facilities for winding bifilar wires, and don't have the super polyimid HV insulated wire required.
The Circlotron runs circles around it. So much easier to wind than the Mac OT besides.
But lets have Pieter make the measurement.
But as I mentioned, we already have the data to decide. The cathode AC voltage and the plate AC voltage. They either sum to twice for one tube, or to zero. Zero means zero gain, no output at all. So they must sum to twice.
If you are getting a gain of 0.75, it just means you need new tubes.
The reason no one can wind one is they don't have facilities for winding bifilar wires, and don't have the super polyimid HV insulated wire required.
Last edited:
Simply attempt to draw a load line and it will become very apparent where the Gain < 1 idea is deficient. If we can't draw a load line, we can't properly design an amp.
No offense but this seems like you're contradicting yourself. Pin 5 of the KT88s measure 149v AC and pins 3 and 8 measure 116v AC. This is not a gain of 1.5, this is a gain of 0.75. 149v AC goes in and 116v AC out. This is a reduction of gain.
You guys must be smoking the same stuff!
Pins 3 and 8 are the plate and cathode. The primary winding is split in two halves, the cathode section and the plate section. So the AC across the full primary is the sum of the two. So (2 x 116) / 149 = 1.55
Its that simple. But please check post 58 to see why.
Pins 3 and 8 are the plate and cathode. The primary winding is split in two halves, the cathode section and the plate section. So the AC across the full primary is the sum of the two. So (2 x 116) / 149 = 1.55
Its that simple. But please check post 58 to see why.
Last edited:
Frank McIntosh's 1949 paper is very clear about the 6L6 version and worth reading. A 4000 Ohm (total) primary has 1000 Ohms per valve split into series'd halves of 250 Ohm each (measured to ground). Each bifilarly wound pair has one valve's anode and the other valve's cathode connected. This has the added benefit of keeping current flowing during the very cold class AB2 operation.
FWIW, the McIntosh circuit can easily be made "Ultralinear" by connecting G2's to the power supply directly, for 50% of turns G2 loading. Factory amps are true pentode because G2's follow cathode voltages.
All good fortune,
Chris
FWIW, the McIntosh circuit can easily be made "Ultralinear" by connecting G2's to the power supply directly, for 50% of turns G2 loading. Factory amps are true pentode because G2's follow cathode voltages.
All good fortune,
Chris
Correct. It's like having regulated voltages for G2 in a conventional PP output stage. G2 is always at the same potential in reference to the tube's cathode. A very clever circuit.
Keep in mind that you can not use the traditional plate curves....
The Cathode/Plate curves need to be derived or measured...
The way to derive the curves is similar to deriving cathode curves...except your subtracting 1/2 the grid 1 voltage from the plate voltage instead of the full Grid voltage..
The other method is to plot the curves on the Tektronix 570 tracer..which I own two of these....
You need the output transformer connected to the tracer to plot the curves..almost same procedure like tracing out Ultra-Linear curves...The plate/cathode curves look almost like cathode curves ...very steep triode looking curves at twice the plate resistance of the follower curves..
Here is why the Gain at the 6550 follower of the MC75 transformer is roughly .75 ... The gm is about 11mA/V ... The cathode "looking in" resistance is roughly 1/gm = 90 Ohms ... Since this is 1/2 a follower the curve slope is x 2 ...now we have 180 Ohms... When you work that into the equivalent load you will see the voltage drop....
When the load line is drawn for a Single Ended stage... it will show a differential output Gain of 2 unloaded similair to the Cathodyne Phase splitter.... This is from a Single Ended input signal source to a Differential output signal.... Properly loaded the Gain is 1.5 ....
BUT the McIntosh Stage is not Single Ended......
It is a differential input signal to a differential output signal...within the PRIMARY windings..
The key factor here is that the differential Push-Pull output voltage signals in the transformer windings do not SUM, since the Push-Pull halves are in PARALLEL.. However, the AC CURRENTS do SUM to contribute to the output power... The Plate and Cathode Winding for each tube will SUM...but the not for both tubes...
Last edited:
For AC, it's like there are 4 separate primaries; 2 of the 4 have 116v across it, they don't add. Now, the way the primaries are physically situated in the transformer, the DC currents cancel. This is why MAC UC amps don't need matched output tubes to optimally operate.
"This is why MAC UC amps don't need matched output tubes to optimally operate."
I think that has to do with the substantial N Fdbk. One can still unbalance the magnetic core if the bias (and avg.) currents are not matched between tubes.
-----------
"Keep in mind that you can not use the traditional plate curves...."
Yes you can. You use a load line equal to the impedance of the plate and cathode windings in series. (2X the turns, so 4X the impedance of each) Then for grid drive, you add 1/2 the total output AC (or the cathode winding portion AC) to the grid swing indicated on the plate curves. Quite straight forward. Ordinary 50% CFB design.
"BUT the McIntosh Stage is not Single Ended......"
Well, any P-P amp is just two single ended amps running in parallel opposed. Class A or class B determines whether the currents actually overlap. (class A overlap then roughly doubling the apparent primary Z)
I think that has to do with the substantial N Fdbk. One can still unbalance the magnetic core if the bias (and avg.) currents are not matched between tubes.
-----------
"Keep in mind that you can not use the traditional plate curves...."
Yes you can. You use a load line equal to the impedance of the plate and cathode windings in series. (2X the turns, so 4X the impedance of each) Then for grid drive, you add 1/2 the total output AC (or the cathode winding portion AC) to the grid swing indicated on the plate curves. Quite straight forward. Ordinary 50% CFB design.
"BUT the McIntosh Stage is not Single Ended......"
Well, any P-P amp is just two single ended amps running in parallel opposed. Class A or class B determines whether the currents actually overlap. (class A overlap then roughly doubling the apparent primary Z)
Last edited:
Of course you can use regular plate curves...and use the plate load x 4 .... not the exact way of showing the feedback effect on the curves that makes them way more linear..
Yes... Those numbers work out if you view this entirely as a Single Ended circuit.....
However, The Mac is a Differential input signal,Grid-to-Grid voltage is 149 x 2.
And your 2 x116 is valid... for each tube Plate and Cathode...
So that's 232/298 = .778
However, that's as far as it goes, the 232V in one winding is in PARALLEL with the 232V in the other Bi-Filar winding and do not SUM these voltages... They do SUM the AC currents in these Parallel Bi-Filar windings... I can put a current probe on the leads and show you...
Last edited:
It's important for us to be clear here: each valve's anode load and cathode load are in series, and both valves' total (anode plus cathode) loads are also in series. All four primary windings are in series and each has 1/4 of total primary voltage across it.
All good fortune,
Chris
I can clarify...
For each valve, the associated 1/2 Plate and 1/2 Cathode windings are in SERIES... this is in PARALLEL with the other half..
All 4 sections are not in series...they can not add up in series based on the phasing as well as both Bi-Filar windings are ground referenced in their centers.. Also this type of amp is not Class A ...actually very close to B and therefore can not cycle current in series...considering the tubes are in cut-off for a significant time..
You will find that the output valves are in series in the same sense that any conventional push-pull amplifier has its output valves in series. It's useful to think of the valves and the two halves of a conventional primary as a bridge, with the primary impedance appropriate for two valves in series.
In the McIntosh circuit all four quarter primaries reference signal ground, so can easily be measured to have 1/4 (each) of required primary voltage for measured output power. Class AB2 operation uses a lower than usual total primary impedance, so may be a confounder.
Frank McIntosh's 1949 paper is in an Audio Anthology I'm pretty sure, and well worth reading. If not available on line, I can scan and post it. Don't think he'll mind.
All good fortune,
Chris
Nice try..but totally of base... How can you run in series when tubes are in cut-off ?? Both tubes need to be ON at all times to circulate a series AC current..Only in Class A..
The primary winding sections are not 1/4 .. I have wound many of these..
I have read and studied the McIntosh papers back when I was a teenager, in fact Sidney Corderman faxed my first copy to me 30 years ago..I have been friends with him for many years and have discussed these transformers with him on many occasion...
The best way to clear this up is simply measure this series connected windings and see if they actually all add up in series.. you will see it wont measure up...
Last edited:
The output transformer of the MC275 has a step down ratio of 16 : 1 for the 9 ohm tap, not 8 : 1.
To verify the voltage gain of the output stage I calculated back, and the outcome perfectly confirms this stage to have a voltage gain of 1.5.
This coincides with the AC voltage chart coming along with the manual.
I am sorry cerrem but in the end you will have to admit this to be a series affair and not a parallel one.
To verify the voltage gain of the output stage I calculated back, and the outcome perfectly confirms this stage to have a voltage gain of 1.5.
This coincides with the AC voltage chart coming along with the manual.
I am sorry cerrem but in the end you will have to admit this to be a series affair and not a parallel one.
That very likely is the reason.
When I am right Gordon Gow stated that actually the unity coupling transformer is easier to wind because of the limited number of primary windings and the better behaviour wrt notch distortion.
Set up a measurement with a certain input voltage (I really mean at the input of the amplifier) and measure the output voltage at the secondary into 8R. Then, leaving everything as is, disconnect the plates from the transformer and connect them to the supply. That's the true pentode cathode follower (yes, you need to drive the screen grids the same way). Now measure again the output voltage and see how much you get. I "guess" this is somewhere between 1.5-2 times lower. As nothing else has changed it follows that the MAC output stage does have a voltage gain regardless of the way you want to look at it.
For the pentode cathode follower stage the driver has to provide a bit more than the full output voltage instead of just about 1/2 of it. It makes all the difference in the world because for high power stuff numbers start to become really high. This is the reason why cathode followers (pentode, triode or ultralinear) are not common otherwise they represent the best kind of output stage in terms of properly driving real reactive loads. The MAC output stage will still behave like a pentode. In particular, if the real load drops and is reactive its driving ability will be poor regardless of its Zout. In fact the power delivery into such lower and reactive load will reduce drastically and distortion will increase a lot. So one cannot use any speaker.....
- Status
- Not open for further replies.