PS: More About Dumbleator Settings

[Stephen1966]

Gil
Sorry for the delay as I got a last minute sub job in DC and the site went out last night.

I didn't have a 15K pot on hand so it was a 50K and 25K pots. Both seemed to work OK. The only thing I noticed was there was a bit more compression when going to the lower value pots. It was pretty subtle so I don't know if you will notice this or not. The really interesting thing was the lower value pots increased the lators ability to sustain (especially the 25K) which was pretty cool!. Of course, the send pots became really sensitive but we already knew this. As much as I tried I couldn't get the CF to distort so it never ran completely out of headroom. (I've heard followers distort)
So far I prefer the 25K for now and am leaving it in there to try on the next few gigs and see how it feels
Of course YMMV depending on your setup tubes/cables and effects.
Master on 7
Lexicon LXP-15
Stock D-Lator
Amps were
High Plate Skyline
Low Plate Skyline Music Man
Let me know if you have any questions.

Tony

Hi Tony, I know I'm late to the show but I've just finished installing a Dumbleator in my Skyliner amp, based on the low plate #124 Skyliner, and I'm a happy bunny

Curious to know what you think... Is it possible the 25k pot increased the 'lators ability to sustain' because of the higher transconductance of the tube with the output impedance a 25k pot offered? I plotted the AC load and it gave about 32V peak which seems to offer plenty of headroom for the signal voltages hitting the grid - a very small risk of clipping.

Stephen

[talbany]

Gil
Sorry for the delay as I got a last minute sub job in DC and the site went out last night.

I didn't have a 15K pot on hand so it was a 50K and 25K pots. Both seemed to work OK. The only thing I noticed was there was a bit more compression when going to the lower value pots. It was pretty subtle so I don't know if you will notice this or not. The really interesting thing was the lower value pots increased the lators ability to sustain (especially the 25K) which was pretty cool!. Of course, the send pots became really sensitive but we already knew this. As much as I tried I couldn't get the CF to distort so it never ran completely out of headroom. (I've heard followers distort)
So far I prefer the 25K for now and am leaving it in there to try on the next few gigs and see how it feels
Of course YMMV depending on your setup tubes/cables and effects.
Master on 7
Lexicon LXP-15
Stock D-Lator
Amps were
High Plate Skyline
Low Plate Skyline Music Man
Let me know if you have any questions.

Tony

Hi Tony, I know I'm late to the show but I've just finished installing a Dumbleator in my Skyliner amp, based on the low plate #124 Skyliner, and I'm a happy bunny

Curious to know what you think... Is it possible the 25k pot increased the 'lators ability to sustain' because of the higher transconductance of the tube with the output impedance a 25k pot offered? I plotted the AC load and it gave about 32V peak which seems to offer plenty of headroom for the signal voltages hitting the grid - a very small risk of clipping.

Stephen

Stephen
I honestly couldn't tell you for certain what's going on and don't have time to run numbers. All I can tell you is what I experienced based on changing out the different pot values. That pot is still in there.
BTW.. Dumble used a 10K send pot in his D-Lator II

https://ampgarage.com/forum/viewtopic.p ... 631#p57631
Good Luck!

Tony

[Stephen1966]

Stephen
I honestly couldn't tell you for certain what's going on and don't have time to run numbers. All I can tell you is what I experienced based on changing out the different pot values. That pot is still in there.
BTW.. Dumble used a 10K send pot in his D-Lator II

https://ampgarage.com/forum/viewtopic.p ... 631#p57631
Good Luck!

Tony

Much appreciated! Encouraging to hear that you also still have the 25k pot in there. I'm quite sure you would have whipped it out by now, if you had had any issues with it. The numbers, look okay. But is it correct that the D'lator II also used a 1uF cap to the send pot as well? Did you change yours along with the pot? Assuming not.

Stephen

[Stephen1966]

Thinking about this some more, and actually thinking, transconductance is near enough a constant regardless of the load placed after it. I calculated it as 1.51mA/V.

What does change is the AC load line. For a 25k pot the Vpeak before saturation and clipping is about 32V, for a 50k pot around 51V and for the 250k pot about 63V, which incidentally is about the same as Martin's buffer mods with a 68k load and 1k5 bias resistor.

Dumbleator CF ECC83S - loadlines.jpg

So, Tony, your 25k pot... I'm thinking the cool sustain is a product of the saturation point being that much closer to the operating point. I didn't run the numbers for a DII but I can imagine that is what MrD was thinking as well with his choice of a 10k pot. The high pass cut off with a 1uF cap and a 10k pot is still below the usable frequency range - about 16Hz.

Installing a 25k pot is my next step along with changing out the 250k Return for a 10k mixer like the Mesa. I might swap out the cap leading to the send pot as well for a larger one because a .47uF cap in combination with the 25k pot is going to affect the low end with a frequency knee at around 136Hz.

I originally thought there was some connection between the transconductance of the tube at its operating point and the sustain, but that's unlikely given it's a constant, unless it is this (gm) in combination with the Vpeak towards saturation. If anyone can offer any suggestions here, I love the numbers and will gladly run them. Of course, ultimately, its more an art than a science and real world measurements aren't likely to tally exactly. Is this some secret sauce though!?

Stephen

[talbany]

Stephen
Thanks for the chart. All of that makes sense. running it closer to saturation might help sustain give you some added compression and 2nd order harmonics . If we think about it acting somewhat like a volume pot (everything else being equal) it would also cut some bass For me the 10k IIRC was too compressed for my taste so I kept the 25K. My send cap is 1uF

BTW. I had problems using a 10k mixing pot on the Mesa mod I think you are talking about doing? Post your schematic so I can take a look at it.

Tony

[Stephen1966]

Stephen
Thanks for the chart. All of that makes sense. running it closer to saturation might help sustain give you some added compression and 2nd order harmonics closer to saturation. If we think about it acting somewhat like a volume pot (everything else being equal) it would also cut some bass which could also help with sustain?. For me the 10k IIRC was too compressed for my taste so I kept the 25K. My send cap is 1uF

BTW. I had problems using a 10k mixing pot on the Mesa mod I think you are talking about doing? Post your schematic so I can take a look at it.

Tony

Tony, sure thing! Thanks. Like a volume pot - hmm. Compression yes, I think the graph shows that easily. The AC loads shown there, all assume 100% on the send pots as well. It didn't factor in what happens to the signal in the tube, when the Send pot is turned throughout its range. (I can feel another graph coming on Another test suggested to me was to put a decade box in place of the send pot and look for the harmonic peaks with the scope with a range of values, 10/25/50k. 25k seems to be the magic number though so maybe 25k/24/23... Just switching through some different values should give us an idea if the affect is small or significant. I can take a look at that. A slightly compressed, harmonically richer top end... hmm. I have to admit, I kinda like the fact that it goes against the grain with the mood for zero-loss/totally transparent loop gizmos.

Here's the circuit I built with (near as damn) measured voltages. This is very well behaved and totally quiet without the bright caps in circuit and on high Send but low Gain settings. It certainly adds a quality of weight and depth especially in the high end.

SKYLINER FX loop - 3.2.jpg

It's got some features that were discussed earlier in this post but it's essentially the Dumbleator you posted in 'Dumbleator Layout' https://ampgarage.com/forum/viewtopic.php?f=5&t=12876

I guess 10k for the Return/Mixer pot was chosen for low output pedals, and 10k promises low noise as well. When its in Series mode, it's effectively acting as a grid leak as well so maybe its better to think of it, and estimate its value from that?

Stephen

[talbany]

The Return/Mixer (in parallel mode) mixing the signal in I found the 250k to be about right. If I went any lower limited my range of wet/dry isolation made for a touchy control and hard to dial in on the fly during gigs. It was also prone to null spots as well (using rack mounted time delays) I tried several different values of linear pots and found a 250k CTS to sound best for me. You shouldn't have this problem in serial mode. Perhaps you might have better luck going with a lower value pot than I did.

Good Luck and let me know what you find!

Tony

[Stephen1966]

The Return/Mixer (in parallel mode) mixing the signal in I found the 250k to be about right. If I went any lower limited my range of wet/dry isolation made for a touchy control and hard to dial in on the fly during gigs. It was also prone to null spots as well (using rack mounted time delays) I tried several different values of linear pots and found a 250k CTS to sound best for me. You shouldn't have this problem in serial mode. Perhaps you might have better luck going with a lower value pot than I did.

Good Luck and let me know what you find!

Tony

I've just reached the same conclusion

Looking up the grid leak suggestions from the datasheet for a typical "anode driver" with a plate resistor of 100k and cathode resistors between 1k5 and 820R, the Rg is 330k with a B+ between 250 and 400V so a 250k return pot is probably about right. I notice that Banza Music in Germany have some 16mm stereo pots and I can probably get one from Mouser... I might go back to that idea. I used a 24mm pot before but it wouldn't fit in this installation and didn't think I could find a 16mm stereo. It works, but using a log100k stereo pot I got a hump in the signal level about the centre position - I think I had it wired back to front! If I use a stereo mixer I think I would use a linear pot this time.

You used a 500k//500k pot is that right! The next size down would be 250k//250k - effectively 125k, maybe too low.

Stephen

[Stephen1966]

To give you an idea of what I'm dealing with in terms of space. The loop circuit is to the right of the lone radial cap.

IMG_20220820_223545.jpg

[Stephen1966]

Another question Tony, if I may...

Do you know if the original Dumbleator(s) had any exceptional power ratings for the load resistor on the CF? I haven't been able to find any information about this but it seems everyone is using ordinary 1/2 or 1 Watt and I think I've spotted at least one RN65D in the wild.

Plate current in my calculations tops out at 9.63mA. If I calculate the power dissipation of the 27k load resistor according to Merlin's advice, I get

0.00963² x 27000 = 2.5W

I'm pretty sure I haven't seen anyone using a resistor that big. I have a 1W in there at the moment - and the house is still standing

Stephen

[pdf64]

…
Plate current in my calculations tops out at 9.63mA. If I calculate the power dissipation of the 27k load resistor according to Merlin's advice, I get

0.00963² x 27000 = 2.5W
…

Surely 9.63mA is the theoretical instantaneous peak current? Such that 180° later, it’s 0mA.
In regard of dissipation, it’s the average value that matters.
But even 4.8mA seems high? 2mA seems more likely.

[Stephen1966]

…
Plate current in my calculations tops out at 9.63mA. If I calculate the power dissipation of the 27k load resistor according to Merlin's advice, I get

0.00963² x 27000 = 2.5W
…

Surely 9.63mA is the theoretical instantaneous peak current? Such that 180° later, it’s 0mA.
In regard of dissipation, it’s the average value that matters.
But even 4.8mA seems high? 2mA seems more likely.

260V/27k = 9.629mA

(0.009629/2)² x 27000 = 0.625W

Agreed, 9.629mA is the theoretical limit and that's actually what seems to be happening... with average value. Course, we could also include the 1k8 of the bias resistor in the Rload as well, but I don't think it would make a lot of difference. The cathode is rated to a maximum 8mA in these tubes in any case. If it really were a problem, this forum would be littered with stories of Dumbleators self-immolating at the first manly power chord

Seriously, the only time when it might reach anything like the theoretical limit of 9mA is at start up in the pico-second where the solid-state rectified charge is unleashed on the plate. Dumbleators have solid state bridge rectifiers as well. However, if anyone has a Q-line 27k flameproof resistor going spare, PM me! Metal films offer precision and they don't appear to drift over time. Carbon films seem better at handling these DC transients though. Correct me if I'm wrong.

[WhopperPlate]

To give you an idea of what I'm dealing with in terms of space. The loop circuit is to the right of the lone radial cap.

IMG_20220820_223545.jpg

I just have to say that this is very inspiring work . I love to just stare !

[Stephen1966]

To give you an idea of what I'm dealing with in terms of space. The loop circuit is to the right of the lone radial cap.

IMG_20220820_223545.jpg

I just have to say that this is very inspiring work . I love to just stare !

Thank you Charlie! What was it, Newton said about standing on the shoulders of giants though!? I've ordered some more components and eventually you will see the finished article. It's so close now.

Stephen

[Stephen1966]

Here's the schematic for the changes I'm planning.

SKYLINER FX loop 3.3.jpg

A 1uF coupling cap along with a 25k "DRIVE" pot to give a cutoff frequency around 6.4 Hz and to keep the output impedance low exploiting the compression for sustain.

Then, a 2 gang linear 500k pot wired antiphase to give a "steady state" 250k grid leak "bias" for the recovery triode. The Series/Parallel switch should work fine in this configuration.