SKYLINER build - inspired by ODS #124 and #060

[ijedouglas]

I would experiment with different PI tubes, although...... if you look at the picture of 183, there is a 120k/100K PI plate resistor combination (as Charlie Wilson pointed out https://ampgarage.com/forum/viewtopic.p ... 80#p400580 ). Mr Dumble probably found a tube he liked and then adjusted the plate resistors.

[Richard1001]

I'm not clear about this step, could you elaborate?


The current is about 45mA unloaded, across the PT cathodes, but where would I set the AC balance

To measure AC balance with the DC voltmeter, you have to put the meter over the cathodes of both tube pairs. You don't measure voltage to ground, but the voltage difference over the cathode resistors (current through the tubes) between both halfs of the poweramp.

So for instance: when the bias on one side is 45ma and on the other side is 46ma, your meter will show this difference (1ma) at zero signal.

[Stephen1966]

I'm not clear about this step, could you elaborate?


The current is about 45mA unloaded, across the PT cathodes, but where would I set the AC balance

To measure AC balance with the DC voltmeter, you have to put the meter over the cathodes of both tube pairs. You don't measure voltage to ground, but the voltage difference over the cathode resistors (current through the tubes) between both halfs of the poweramp.

So for instance: when the bias on one side is 45ma and on the other side is 46ma, your meter will show this difference (1ma) at zero signal.

Oh I get it, yes, thank you, that makes sense... and I have a difference of about 2mA between one side of the OT and the other. The +2mV/-2mV swing I see, it all makes sense.

[Stephen1966]

On the subject of the input frequency, 400Hz corresponds roughly to (fourth fret) G#, 800Hz to (sixteenth fret) G# to and 1kHz to just shy of the B (19th fret) - all on the first string. These are also roughly speaking, the positions of upper partials on the instrument. I am just wondering if there is some reason we are calibrating the amp so close to the natural upper partials. It curious, but I see a pattern there that might be exploited. 329 Hz (the fundamental, E4) 440Hz (A4), 494Hz (B4), 659Hz (the octave, E5), 880Hz (A5), 988Hz (B5) and 1319Hz (the second octave, E6) might all be useful as reference points for the second and third upper partials off the fundamental. All very well, if we play the guitar in standard tuning. But what if we play in Eb standard tuning, or dropped D??? I guess, just use the fundamental, sub-dominant and dominant frequencies of those tunings... I don't know if any of this holds water, but it doesn't seem accidental that you chose test frequencies so close to the actual tuning of the instrument.

A digression perhaps...

On the other hand, I can see the sense of using a higher frequency to calibrate with. Since with a higher frequency the DC ripple element is proportionally smaller with the higher the frequency. My test frequency used in the above was 1022Hz as you see in the scope. Same frequency with the cathode part of the test.

[Stephen1966]

I would experiment with different PI tubes, although...... if you look at the picture of 183, there is a 120k/100K PI plate resistor combination (as Charlie Wilson pointed out https://ampgarage.com/forum/viewtopic.p ... 80#p400580 ). Mr Dumble probably found a tube he liked and then adjusted the plate resistors.

I think you are probably right, and the schematics point to a 10k trimmer there. I don't think 120k and 100k seems so extreme, with the 10k trimmer it still adds up to something close to the external resistance of the 124's 235K. In Tony's first picture https://ampgarage.com/forum/viewtopic.p ... 53#p400553 there is a 4k7 cap as well which doesn't appear in the schematics I've seen. Any idea what that is there for? It doesn't appear to have been improvised.

[ijedouglas]

I would experiment with different PI tubes, although...... if you look at the picture of 183, there is a 120k/100K PI plate resistor combination (as Charlie Wilson pointed out https://ampgarage.com/forum/viewtopic.p ... 80#p400580 ). Mr Dumble probably found a tube he liked and then adjusted the plate resistors.

I think you are probably right, and the schematics point to a 10k trimmer there. I don't think 120k and 100k seems so extreme, with the 10k trimmer it still adds up to something close to the external resistance of the 124's 235K. In Tony's first picture https://ampgarage.com/forum/viewtopic.p ... 53#p400553 there is a 4k7 cap as well which doesn't appear in the schematics I've seen. Any idea what that is there for? It doesn't appear to have been improvised.

I believe the 47pf cap is a snubber which you can find on the BluesMaster layout/schematic

[Stephen1966]

I would experiment with different PI tubes, although...... if you look at the picture of 183, there is a 120k/100K PI plate resistor combination (as Charlie Wilson pointed out https://ampgarage.com/forum/viewtopic.p ... 80#p400580 ). Mr Dumble probably found a tube he liked and then adjusted the plate resistors.

I think you are probably right, and the schematics point to a 10k trimmer there. I don't think 120k and 100k seems so extreme, with the 10k trimmer it still adds up to something close to the external resistance of the 124's 235K. In Tony's first picture https://ampgarage.com/forum/viewtopic.p ... 53#p400553 there is a 4k7 cap as well which doesn't appear in the schematics I've seen. Any idea what that is there for? It doesn't appear to have been improvised.

I believe the 47pf cap is a snubber which you can find on the BluesMaster layout/schematic

Forgive me if I have this wrong. I see 4k7, don't you mean 4700pF?

[ijedouglas]

Nope. Check out the BluesMaster https://ampgarage.com/forum/viewtopic.php?f=5&t=12875

[Stephen1966]

Nope. Check out the BluesMaster https://ampgarage.com/forum/viewtopic.php?f=5&t=12875

Thanks, I see 47pF on the layout. I see others of the 183 that don't use it as well so I assume it isn't a critical addition. Seems like a good idea (as a snubber) though.

Edit: Looking at it again the photo actually shows 47K so I was wrong on both counts

Structo's schematic also shows clearly 47p(F): https://ampgarage.com/forum/viewtopic.p ... 62#p223262 I guess the K in 47K is the tolerance code, 10%.

[Richard1001]

I would experiment with different PI tubes, although...... if you look at the picture of 183, there is a 120k/100K PI plate resistor combination (as Charlie Wilson pointed out https://ampgarage.com/forum/viewtopic.p ... 80#p400580 ). Mr Dumble probably found a tube he liked and then adjusted the plate resistors.

I think you are probably right, and the schematics point to a 10k trimmer there. I don't think 120k and 100k seems so extreme,

When you calculate the amplification of both sides of the long tail PI, given the value of the resistors used on the tail and powertube grids,(820 ohm, 24k, 220k) the anode resistors should be +/- 10k apart (91 and 100k or 110 and 110k) to get close to the same gain on both sides with a balanced tube. If you then add a 10k variable resistor to feed the anode resistors, you can set the balance easy. Even if the tube itself is not balanced.

With a difference of 20K between anode resistors you allready have to much unbalance for most tubes to set the balance with a 10k pot. So if that's the case, either use a bigger value pot or get the anode resistors closer in value.

[Stephen1966]

I would experiment with different PI tubes, although...... if you look at the picture of 183, there is a 120k/100K PI plate resistor combination (as Charlie Wilson pointed out https://ampgarage.com/forum/viewtopic.p ... 80#p400580 ). Mr Dumble probably found a tube he liked and then adjusted the plate resistors.

I think you are probably right, and the schematics point to a 10k trimmer there. I don't think 120k and 100k seems so extreme,

When you calculate the amplification of both sides of the long tail PI, given the value of the resistors used on the tail and powertube grids,(820 ohm, 24k, 220k) the anode resistors should be +/- 10k apart (91 and 100k or 110 and 110k) to get close to the same gain on both sides with a balanced tube. If you then add a 10k variable resistor to feed the anode resistors, you can set the balance easy. Even if the tube itself is not balanced.

With a difference of 20K between anode resistors you allready have to much unbalance for most tubes to set the balance with a 10k pot. So if that's the case, either use a bigger value pot or get the anode resistors closer in value.

The 91k / 10k / 100k combination seems like a safe bet. A 10% difference is probably what most people would recommend here but Aiken and I think, Blencowe mentioned it could be anything up to 20% smaller on the input (inverting side) - when there is only one channel and a negative feedback. The Bassman, part of Blencowe's study used an 82k / 100k (no pot), though with just two output tubes. Essentially the same design of power amp. I think I have a 100k and a 10k pot somewhere. I could try the pot first and then if that doesn't cut it, install a 100k on the feedback side and find the other (smaller) resistor, using the decade box. It just feels like 5k is like trying to drive fast in first gear - getting nowhere, fast.

[Richard1001]

I think you are probably right, and the schematics point to a 10k trimmer there. I don't think 120k and 100k seems so extreme,

When you calculate the amplification of both sides of the long tail PI, given the value of the resistors used on the tail and powertube grids,(820 ohm, 24k, 220k) the anode resistors should be +/- 10k apart (91 and 100k or 110 and 110k) to get close to the same gain on both sides with a balanced tube. If you then add a 10k variable resistor to feed the anode resistors, you can set the balance easy. Even if the tube itself is not balanced.

With a difference of 20K between anode resistors you allready have to much unbalance for most tubes to set the balance with a 10k pot. So if that's the case, either use a bigger value pot or get the anode resistors closer in value.

The 91k / 10k / 100k combination seems like a safe bet. A 10% difference is probably what most people would recommend here but Aiken and I think, Blencowe mentioned it could be anything up to 20% smaller on the input (inverting side) - when there is only one channel and a negative feedback. The Bassman, part of Blencowe's study used an 82k / 100k (no pot), though with just two output tubes. Essentially the same design of power amp. I think I have a 100k and a 10k pot somewhere. I could try the pot first and then if that doesn't cut it, install a 100k on the feedback side and find the other (smaller) resistor, using the decade box. It just feels like 5k is like trying to drive fast in first gear - getting nowhere, fast.

The bigger the tail resistor, the more balanced the output will be. And the lower the tail resistor...

The Fender bassman uses 82/100k but also a 10k tail resistor.

[Stephen1966]

...
I believe the 47pf cap is a snubber which you can find on the BluesMaster layout/schematic

I've been thinking (dangerous, I know) of the purpose of a snubber between the two sides of the PI. If it is snubbing parasitic inductance, where is that most likely to occur...? Is this a case of mutual inductance with the two plate resistors so close together? In this case, is there a case to be made for using metal film resistors? I understand MFs designed specifically for high frequencies, are better for their inductance properties than say, wirewound (not that you would normally use WW in that position). But if it is mutual inductance, and it went untreated, I also understand that this is likely to introduce noise at high frequencies in the AC signal. The cap, may not be necessary, but it is a nice touch if I understand the problem correctly. It reminds me of MrD's emphasis on observing circuit constants, and seems a conscious attempt to keep the signal as smooth as possible heading from the PI. And maybe, also, there is method here of taming inductance in the tube as well...

I found an interesting paper on the subject of tubes at high frequencies which talks about 'Lead inductance and interelectrode capacitance', though the work seems to be concentrating on the efficiency of tubes working at ultra high frequencies (television bandwidth region and higher). I'm not sure this is entriely relevant then, but here's the link... (page 44)
https://www.google.co.uk/url?sa=t&rct=j ... o0TWwHxR3x

Another paper, talks about when the load is inductive there being a regenerative effect on oscillation... (for which, I read "noise")
https://www.google.co.uk/url?sa=t&rct=j ... rxEoPu2UZZ

I see more interesting aspects of the problem of input impedence in this last paper as well, but I'm just trying to understand the thinking behind that capacitor.

[talbany]

It reminds me of MrD's emphasis on observing circuit constants, and seems a conscious attempt to keep the signal as smooth as possible heading from the PI. And maybe, also, there is method here of taming inductance in the tube as well...

Stevphen
AFAIK Dumble used the cap across the plates of the PI on his Bluesmasters only, (aka Marshall/Bassman) Why?
My impression is? Same reason Fender/Marshall used it.
The combination of both the 10k tail and 511 RK as well as a reduction in GNFB, gives you a hotter PI,with more gain harmonically layered w/ less dampening, therefore more susceptible to H.F Oscillation's at peak signal levels vs the better balanced lower gain versions used in the in the later 4th generations, which is going to be a smoother/cleaner/ Flatter/Tighter sound. I doubt it's because he used different resistor types? (tone shaping aside)
You are right, since the plates are out of phase your simply canceling (NFB) to those (depending on the size of the cap) frequency's, giving a smoother less aggressive type sound, I've never seen one on a 4th generation amp given it was already plenty smooth w/ symmetrical output and have yet had one oscillate on me..If you want to go to an even smoother cleaner punchier PI use the one from a 2nd Generation w/ AT7

BTW.I think it was David Hafler that recommended using non inductive resistors in the RK of the PI in his amps, not sure on the RP side. I will try to find that paper
Tony

[Stephen1966]

It reminds me of MrD's emphasis on observing circuit constants, and seems a conscious attempt to keep the signal as smooth as possible heading from the PI. And maybe, also, there is method here of taming inductance in the tube as well...

Stevphen
AFAIK Dumble used the cap across the plates of the PI on his Bluesmasters only, (aka Marshall/Bassman) Why?
My impression is? Same reason Fender/Marshall used it.
The combination of both the 10k tail and 511 RK as well as a reduction in GNFB, gives you a hotter PI,with more gain harmonically layered w/ less dampening, therefore more susceptible to H.F Oscillation's at peak signal levels vs the better balanced lower gain versions used in the in the later 4th generations, which is going to be a smoother/cleaner/ Flatter/Tighter sound. I doubt it's because he used different resistor types? (tone shaping aside)
You are right, since the plates are out of phase your simply canceling (NFB) to those (depending on the size of the cap) frequency's, giving a smoother less aggressive type sound, I've never seen one on a 4th generation amp given it was already plenty smooth w/ symmetrical output and have yet had one oscillate on me..If you want to go to an even smoother cleaner punchier PI use the one from a 2nd Generation w/ AT7

BTW.I think it was David Hafler that recommended using non inductive resistors in the RK of the PI in his amps, not sure on the RP side. I will try to find that paper
Tony

What about the following impedance of the PT grid leak resistors? 100k in the second generation ODS 50W https://ampgarage.com/forum/viewtopic.p ... 59#p140259
Could we leave those at 220k as in the later generations?

Second generation... I do declare



[Edit: Tony informs me, this amp is "David Lindley's old amp #002. It's a variant on a 2nd generation..."]

I guess it doesn't hurt to use non-inductive resistors on the plates as well. I'm considering using carbon film, for their low impedance/low inductance properties, though their tendency to drift as they dissipate heat might negate any benefit. So, metal film might be a better choice there. To be clear, I don't hear or see any oscillation either. With the reverb Return turned down, the amp is dead quiet. The transformers make more sound in the chassis than the unloaded signal does in the speaker. And when the volume is up and hitting notes high up the neck, it's very tight and smooth... and articulate: the attack portion of the note is enough to require ear defenders. I would call it "punchy" - it actually hurts. Pumping a sine wave in at high levels and high frequency (around 1kHz) results in a clean sine wave at the output end - at the PI plates and further on at the PT cathodes.

I think this is a good place to start with the Bluesmaster:
https://ampgarage.com/forum/viewtopic.php?f=5&t=12875

It's interesting to note that the 110k Rp (on the non-inverting feedback side) is paired with a 91k Rp (on the input side) along with the 511R (Rk) and 9.09k (tail). The NFB resistor values here are much, much larger than in the 4th generation. This suggests a very "edgy" tone.

Something like this (5th generation ODS) perhaps...?



This is a sound I can dig, but being 5th generation, if I'm correct, I would expect the PI to be set up like the #183 https://tubeamparchive.com/viewtopic.ph ... 00#p143400.

This isn't the first time I've deferred to that design - my Bright switch is cobbled together from the 183 (perhaps, I should rename this topic to reflect that... ah, well!)

I just think the 110k Rp / 5k (trimmer) / 120k Rp, in the #124 is just a bit sterile for me. I don't know if changing these to the 183 values (100k/10k/110k) will boost the gain any noticably as you seem to suggest it is the cathode resistor which has more impact here, but if I can get that ~7V imbalance in the plates, I'm sure I can call up the second harmonics and access that elusive bloom. There is already a complexity to the note and it sustains well with clear upper partials emerging after the fundamental has decayed - I just feel this amp has more to offer and I just have to find it.

I am humbled by the sheer breadth of knowledge of you guys. A beacon of light in the basement of amp construction.

Thank you.