Q about PI trimmer design

[Structo]

The trimmer does adjust the gain of the PI tube but most have said that it usually ends up where the input has about 6v higher voltage.

I have quite a few notes about this.
Henry wrote:
You use 100k on the input plate and 110k on the GFB plate and adjust to make the input plate 3 to 6 volts higher. Listen for accented 2nd harmonics (octaves) to adjust.

dogears wrote:
There is no way to tell a correct setting. I can say that for me, and when using a nicely balanced PI tube, it is usually about 12K or 13K between the plates. I usually wind up with 5-8V more on the first plate.

Gil wrote:
I have to say that my experiences in this department have been very varied. In one amp, the juciest harmonics would occur when the non feedback side had the higher voltage.

For what it's worth, I always get a higher voltage reading on the 110K, which is the input side (110K) and pin 6 on my amp and on Dumbles.
I have never had it where the feedback side was higher.

I use the 110K/ 120K with 10K trimmer.

But the best way to adjust the trimmer is to have the amp quite loud and brush the strings lightly while you adjust the trimmer.
Be very carefully here and use an insulated screw driver.

Make small adjustments and wait for the change to take full effect.
It takes a second or two to take effect.

When you get the right spot, the guitar will want to bloom into feedback and the overtones will be much more pronounced.

[dogears]

One could use a 100K with 20K variable resistor in series on the first triode and a 120K on the second. End result, although arrived at in an unorthodox manner, is identical.

Also, regarding Structo's post above, I always have more voltage on the non feedback first triode. But, I always used perfectly matched and balanced (for both gain and transconductance) tubes. I have seen great sounding tubes have more voltage on the feedback side, sure, just not if they are perfectly balanced.

[Max]

I have quite a few notes about this.

Henry wrote:
You use 100k on the input plate and 110k on the GFB plate and adjust to make the input plate 3 to 6 volts higher. Listen for accented 2nd harmonics (octaves) to adjust.

dogears wrote:
There is no way to tell a correct setting. I can say that for me, and when using a nicely balanced PI tube, it is usually about 12K or 13K between the plates. I usually wind up with 5-8V more on the first plate.

Gil wrote:
I have to say that my experiences in this department have been very varied. In one amp, the juciest harmonics would occur when the non feedback side had the higher voltage.

Hi Structo,

Alexander Dumble recommended to adjust these "dynamic balance" trim pots of his amps (external - like in the ODS 150W and some DL 300SL - or internal - like in all ODS amp since the transition generation "classic" ODS amps) for lowest THD at around half the maximum output voltage just as it is explained in principle in the SVT service manual and power amp schematic (see the description of the "calibration procedure" in the attched SVT power amp schematic).

The idea is according to Alexander Dumble to adjust the "dynamic balance" trim pots for maximum power amp symmetry (or dynamic balance) over as broad a range of output voltages from low to high as possible. The better the power amp tubes are matched not only for idle current but for also for transconductance, the better the symmetry will be IMO in a broad range of output voltages.

Cheers,

Max

[greiswig]

Alexander Dumble recommended to adjust these "dynamic balance" trim pots of his amps (external - like in the ODS 150W and some DL 300SL - or internal - like in all ODS amp since the transition generation "classic" ODS amps) for lowest THD at around half the maximum output voltage just as it is explained in principle in the SVT service manual and power amp schematic (see the description of the "calibration procedure" in the attched SVT power amp schematic).

Max,

I've seen that schematic before, but the copy is always so poor it is hard to read. So here is my transcript of what I think it says, with ??? marking where it is particularly ambiguous:

1. Connect a 4 ohm load resistor to pins 1 and 4 and short pins 2 and 3 of connector P3 (X, Y)

2. Output stage bias current adjustment. (The bias can be adjusted by removing the head cabinet baffle. Gain access from the rear of the cabinet by using the space available through the cooling fan blades)
a. Adjust VR1 for +/- 0.72V (???) DC between K1 and ground
b. Adjust VR2 for zero +/- .01 volts between K1 and K2

3. Phase inverter balance control adjustment
a. Harmonic Distortion Meter Method
i. Drive amplifier to 25V RMS out at 40Hz and connect distortion meter to load resistor. Adjust VR3 (???) for minimum distortion.
b. Volt Meter Method
i. Adjust output as in step 3a, and connect DC volt meter between testing points K1 and K2. Adjust VR3 for zero +/- .01V (???)
ii. The load resistor should be able to dissipate the full power of the amplifier, or 330W.

K1 and K2 are the plates of the respective halves of the push-pull output section. So the meat of this seems to be to adjust this with the power amp pushed fairly hard into a dummy load, and to try to get the plates on the power tubes to come to 0VDC between them?

I've often wondered about this method, since I struggle to find the "sweet spot" in the trimmer. If I could get to the balance point first, then start adjusting slightly away from it, I was hoping I might have better success at finding it.

[greiswig]

One could use a 100K with 20K variable resistor in series on the first triode and a 120K on the second. End result, although arrived at in an unorthodox manner, is identical.

Also, regarding Structo's post above, I always have more voltage on the non feedback first triode. But, I always used perfectly matched and balanced (for both gain and transconductance) tubes. I have seen great sounding tubes have more voltage on the feedback side, sure, just not if they are perfectly balanced.

Thank you, Mr. Lerner!

[dogears]

Nice! I am an anal retentive bastard when it comes to matching my outputs for both current and transconductance. I always felt I got more girth and bloom that way!!!

The idea is according to Alexander Dumble to adjust the "dynamic balance" trim pots for maximum power amp symmetry (or dynamic balance) over as broad a range of output voltages from low to high as possible. The better the power amp tubes are matched not only for idle current but for also for transconductance, the better the symmetry will be IMO in a broad range of output voltages.

Cheers,

Max

[ayan]

Hi George,

I don't know what to suggest, other than you try to install additional eyelets and do things the conventional way... I have an idea for you if you don't want to drill the board: you can etch a little PCB (as I have done in my amps for the treble bleed circuit at the output of the overdrive). You will simply mount the trimmer to it, and the PC board will connect to the main board via three wires (solid core, so they act as stand offs): B+ to wiper, and one connection each to the end of the plate load resistors.

Not sure if the above helps, but it might.

Cheers,

Gil

[ayan]

Hi George,

I don't know what to suggest, other than you try to install additional eyelets and do things the conventional way... I have an idea for you if you don't want to drill the board: you can etch a little PCB (as I have done in my amps for the treble bleed circuit at the output of the overdrive). You will simply mount the trimmer to it, and the PC board will connect to the main board via three wires (solid core, so they act as stand offs): B+ to wiper, and one connection each to the end of the plate load resistors.

Not sure if the above helps, but it might.

Cheers,

Gil

PS: Obviously, the plate load resistors will have to mount on the PCB as well....

Gil

[ampdork]

Get pop riveter, (harbor freight)
Drill for eyes,
remove rivet from stake,
replace with eyelet,
rivet in eyelets,
tap out left over tidbits,
hook it up,
easier then it sounds really.
thank Henry (heistl) for the useful trick.

[Structo]

That is a neat trick thanks for posting it.

Angelo made a little eyelet board for his amp and secured it to the corner bolt of the board.

[greiswig]

Thanks, Gil. I did it a different way, one that I've been thinking about anyway.

I got a 10k pot and mounted it on the rear of the chassis, so it can be adjusted without pulling the chassis out of the cab. I have flying leads from the board to the pot, resistors mounted to the board. It should work.

I don't notice a difference yet in being able to find the sweet spot...still elusive.

Thanks for the suggestions, everybody! And thanks for explaining some of the principles.

Hi George,

I don't know what to suggest, other than you try to install additional eyelets and do things the conventional way... I have an idea for you if you don't want to drill the board: you can etch a little PCB (as I have done in my amps for the treble bleed circuit at the output of the overdrive). You will simply mount the trimmer to it, and the PC board will connect to the main board via three wires (solid core, so they act as stand offs): B+ to wiper, and one connection each to the end of the plate load resistors.

Not sure if the above helps, but it might.

Cheers,

Gil

[Max]

I've seen that schematic before, but the copy is always so poor it is hard to read. So here is my transcript of what I think it says, with ??? marking where it is particularly ambiguous:

OK, step by step:

1. Connect a 4 ohm load resistor to pins 1 and 4 and short pins 2 and 3 of connector P3 (X, Y)

The kind of load resistor of course depends on the maximum output and the OT specs (4,8,16 ohm) of the amp you want to calibrate. AFAIU the SVT schematic the test signal is applied to the power amp input.

2. Output stage bias current adjustment. (The bias can be adjusted by removing the head cabinet baffle. Gain access from the rear of the cabinet by using the space available through the cooling fan blades)
a. Adjust VR1 for +/- 0.72V DC between K1 and ground

These test points of an SVT make it possible (1 ohm resistor) to measure the idle current of three of the output tubes (pull or push section) as the DC voltage between K1 and ground (G) or K2 and ground (G).

VR (variable resistor) 1 and VR2 are the bias adjustment pots and VR3 is what Alexander Dumble calls the "dynamic balance" pot.

b. Adjust VR2 for zero +/- .01 volts between K1 and K2

If the voltage between K1 and K2 measures close to zero then the idle current of the push and pull section of the output tubes of the SVT is exactly the same. Because the idle current often drifts a little bit, this method is thought to be more precise than first adjusting K1 > G for the specified idle current (DC voltage) and then K2 > G for the same idle current (DC voltage).

3. Phase inverter balance control adjustment
a. Harmonic Distortion Meter Method
i. Drive amplifier to 25V RMS out at 40Hz and connect distortion meter to load resistor. Adjust VR3 (???) for minimum distortion.

The 25V RMS are of course only correct for the SVT. I did not calculate it again right now but AFAIR these 25V RMS are around the half of the specified maximum output of an SVT.

This minimum THD method has the advantage that by this method you can directly measure the result you are after in all this procedure of adjusting the dynamic balance for maximum symmetry (and including the effects of the influences of the OT as an example): minimum THD of the power amp section.

The idea is to think of the preamp as the "tone generator stage" and the power amp as the famous "wire with gain" that should only amplify the signal present at the preamp output but not alter it in any way.

b. Volt Meter Method
i. Adjust output as in step 3a, and connect DC volt meter between testing points K1 and K2. Adjust VR3 for zero +/- .01V (???)

The idea is that if you measure a zero voltage between K1 and K2 at - at least - two points of the power amp operation - when idling and at around half of the maximum output - that then the push and pull sections will work with sufficient symmetry to achieve a minimum THD value in the complete range of the power amp operation.

ii. The load resistor should be able to dissipate the full power of the amplifier, or 330W.

The power dissipation ability depends of course on the output power of the amp you want to calibrate.

Cheers,

Max

[greiswig]

Max,

Fantastic explanation! Very helpful. Thank you for taking the time.

[Max]

Hi greiswig,

if you and/or others should want to calibrate amps by this method which don't have these 1 ohm resistor test points you could perhaps use devices like the TAD "Bias Master" (if you know how to connect the sockets to your multimeter to measure the DC voltage, you will perhaps only have to buy the number of sockets you want), wich could perhaps be built by an experienced DIY builder, too: http://www.tubeampdoctor.com/index.php? ... anguage=en

A load resistor that does not damage the amp could probably be built by most DIY builders, too.

And the load resistor could of course be replaced by a power attenuator (if you should like to have one anyway) that has a "Load" option and because of this for sure will be able to take the full load even when not connected to a speaker, like the TAD Silencer as an example: http://www.tubeampdoctor.com/product_in ... anguage=en And a DIY builder could perhaps build something like this, too.

And a simple test signal generator could perhaps be bought used without spending a fortune.

Have a great weekend,

Max

[ayan]

I don't notice a difference yet in being able to find the sweet spot...still elusive.

Try another PI tube, I have gone through a few before I find one that delivers the goods.

Gil