Ultralinear OT and available voltage swing

[priesty]

While mulling over some tube data, I noticed that using the "dreaded" Ultralinear on a pair of 6550's or KT88's can handle a greater peak to peak input signal, compared with tetrode connection at similar B+.

Now If I have a DIY bass amp that is clipping symetrically at moderate volume levels (i.e. less than half way turned up) would ultralinear be a good solution to make more of the upper reaches of the power more usable for bass guitar?

I reckon I'm getting 30 to 40W (based on scope readings) undistorted in tetrode connection before clipping sets in. I estimate that with a 425V B+ I would get 60W from a pair of KT88 into a 4.3K Hammond 1650N.

The taps are there unconnected at the moment, so I suppose it would be an easy mod to try......

[Enzo]

Yes, I think that reporting to us what happens is more valuable than asking us what we THINK will happen. Change two connections and you are there.

Without me studying the curves, did your research show greater output for greater grid swing? Or did it show the need for greater grid swing to get to the same power?

I would also suggest if your amp is clipping, find out where. If your PI is clipping, for example, your power stage will dutifully try to amplify that clipped signal.

[MWJB]

"I would also suggest if your amp is clipping, find out where. If your PI is clipping, for example, your power stage will dutifully try to amplify that clipped signal." ...good call. as your B+ is relatively low for a 6550/KT88 amp (and therefore all later voltages) & there's not much you can do about that without changing the PT, you might explore raising the voltages to the PI & preamp tubes by reducing the value of the preamp dropping resistors? Try the BF values of 1K between B+/screens & PI, then 4.7K between PI & preamp.

If voltages go too high the amp may sound harsh, but 250-270vdc on a 12AX7 in V1 might be OK (or <190vdc on a 12AY7 in V1)?

[wizard333]

A completely un-technical answer, but, apparently FMIC thought that the answer was yes since they upped the claimed power ratings when they switched to ultra linear design. I'm guessing they had a solid reason for that other than marketing.

[JoeM]

A completely un-technical answer, but, apparently FMIC thought that the answer was yes since they upped the claimed power ratings when they switched to ultra linear design. I'm guessing they had a solid reason for that other than marketing.

It wasn't marketing. The B+ was somewhere around 500V on the SFSR (UL version) I had. One loud and clean amp, like a Twin with 4X10s.

[priesty]

Enzo - I suppose you are right, I should just try it and report my findings! I'll maybe get time to work on it between now and New Year.

As for pre-amp or PI clipping, I scoped the PI output, and after improving the balance by "lengthning" the long tail pair from 10K to 47K (and improving the balance of the plate resistors to suit) the signal hitting pin 5 of the power tubes is clean as can be.

I note your comments about low B+, just wondering if the ultralinear lets me make the most of the B+ I have.

If I do try this mod, would I take the B+ from the "smooth" (i.e. screen) side of my filter choke?

[ThermionicScott]

If I do try this mod, would I take the B+ from the "smooth" (i.e. screen) side of my filter choke?

Only if your choke is rated ALL of the plate current your amp will use. I wouldn't use a blackface-style 40mA or 90mA choke for this. The tweed Fenders that did this had a 200mA (or something like that) choke.

- Scott

[MWJB]

"As for pre-amp or PI clipping, I scoped the PI output, and after improving the balance by "lengthning" the long tail pair from 10K to 47K (and improving the balance of the plate resistors to suit) the signal hitting pin 5 of the power tubes is clean as can be." Such a large PI tail will surely be knocking down the gain (assuming that you have global NFB), 10K, or maybe even 6.8K (with proper balance) would surely be better for driving the KT88's?

[Steve Conner]

I'd also recommend checking the screen voltage and screen resistor values. If the screen voltage is too low or the resistors are too big, the tubes won't make the power.

One of my homemade amps had about the same B+ and was originally meant for 50W with EL34s. Using a 1k screen resistor shared between the two tubes, and a mismatch of one step down (so the tubes saw 3.3k a-a instead of the 6.6k design load of the OT) I got about 60-70W with 6550s or KT88s. Output was limited by the power supply, which would sag down below 400V. Previously it had been limited by the EL34s clipping at their grids.

My motivation was that the OT only had one output tap, 16 ohms, and I wanted to make the amp usable with 8 ohm loads.

UL won't increase the power output, except in so far as it makes more voltage available to the screens, by getting rid of the screen resistors.

[priesty]

Wow - thanks for the advice guys! In answer to your points:

ThermionicScott: Good point about the choke - its a Hammond 193N (3H @ 500mA) so should be OK.

MWJB: I have lots of gain (12AY7, 12AX7 and 12AX7 PI which I have changed to a 12AT7) so I took Morgan Jones Book's route of making the Long Tail resistor large and more "Current Sink" like to improve my balance. The remaining calculated imbalance with the 12AT7 lead me to change the 82K and 100K plate resistors to 120K and 130K. Even with this reduced gain, I still have more peak to peak voltage swing than the KT88's can handle, hence why I was mulling over ultralinear.

Steve: OT Primary of 3.3K would get a set of KT88's singing - right enough. I have maybe been a bit conservative with a 4.3K unit with my limited B+. I have a small representation of the power section attached, hope it makes sense (2 x 1K screen resistors).

[ThermionicScott]

6W is way overkill for those 1-ohm cathode resistors, but won't hurt anything.

- Scott

[J M Fahey]

Hi Priesty.

I reckon I'm getting 30 to 40W (based on scope readings)

Your "window" is too large, obscuring whatever power gain you may have going UL.
I'm a firm believer in modifying *one* variable at a time, measure, and then another one and so on, or you can't be sure about anything.
Ah!, and starting with an unmodded amp.
If your main question is about UL, I'd use the *original* schematic, that means original PI, +B connection to the original (capacitive) point, use the choke only for screen, etc. *and* test it with screen grids to the original point with resistors, measure, then UL connected, measure.
To have more precision than counting lines on the scope, use a 400Hz signal and measure the voltage with your multimeter.
Most general purpose multimeters can be trusted up to 400Hz, and even if they are not flat, the loss will be the same for both measurements, so any comparison will still be valid.
You do not even need a True RMS voltmeter in this case.
Regular multimeters really measure *average* or *mean* voltage, but the scale is calibrated in RMS voltage.
The result is valid and accurate *for sine waves*, so if you drive your amp to visible clipping and then back up until it just disappears, fine.
I'm certain you will have *some* improvement, probably not more than 5 W RMS.
Please test and post, your experiment is interesting.
After that, you can vary secondary load to vary reflected primary impedance, and test power output in each case, you could post a table of power vs. impedance, and even draw a graph, if you wish.
You can also calibrate your multimeter's frequency response, if you wish.

[priesty]

Just an update on the above thead to complete the story for anyone researching Ultralinear in Bass amps.

I took the B+ from the "downstream" side of the Hammond 193N, making the power supply for OT centre tap a 50uF-3H-50uF PI filter. Voltage regulation has improved as a result; when running against a signal generator at max volume B+ only drops 3 volts. I have made the output 6550 screens switchable between Pentode and Ultralinear. I get 55 watts RMS from both (with appropriate adjustment of bias) although for Bass use the UL mode appears louder. This was measured at 250Hz 0.5V input signal against a 7.5ohm dummy load. Screen resistors 1K 5W.

In use, I had first band practice with the amp in it's new configuration last night. I was running through an Ampeg SVT 15E cabinet rated at 200W @ 98dB (1x15" ported). The amp was in Ultralinear mode at hottish bias of 70mA per tube. Playing classic rock/metal (Thin Lizzy, Iron Maiden, Black Sabbath) with a loud drummer and two guitarist playing through 60W Marshall combos, the amp was cranked to the max on the volume and could more than keep up volume wise. The clarity and tightness of bottom end available with the UL and ported 15" speaker were superior to any bass amp I've played before (if I don't mind saying so myself!!). I will hopefully have clips to go on YouTube soon.