Why would the frequency response vary so much on a pair of Monoblock amps?

[sgerus]

I’ve been testing the frequency response on my
EF37A/2A3-40 amps and noticed the high end frequency response is quite different from the left amp and the right amp.

Left, 36KHz@-3db
Right, 57KHz@-3db

The EF37A is wired as a triode, C-L coupled to the 2A3

I’m using the same tubes (EF37A and 2A3) when testing, and the same settings on the scope.

Both amps are about the same at the low end. The scope looks good down to 10hz. That seems to be as low as I can measure with a 60Mhz scope?

Are these numbers realistic for this type of tube amp?

Any idea why the frequency response could be so different?

 

[korneluk]

Process of elimination - try switching the tubes between monoblocks one at a time, first input, and then output, and see if it follows any one tube. If not, then the variability must reside in the OPTs and other components in the signal chain.

-- josé k.

 

[kevinkr]

Hi Scott,
There could be multiple reasons, but I would start by measuring the bandwidth of each stage in both amplifiers and doing a comparison.. Much better than guessing.

It could be a difference in the construction of the two plate load chokes, the orientation of the "hot" and "cold" ends of the chokes relative to each other, a difference in the opts (A resonance in one bringing up the HF response or higher leakage inductance and/or interwinding capacitance of one relative to the other) or a highly inductive cathode bypass capacitor in the output stage of the channel with the poorer response. (I had this problem with some unknown brand PPIO motor runs I used as cathode bypass caps in an amplifier.)

IMO It is significant enough that I would attempt to track it down.

 

[sgerus]

It could be a difference in the construction of the two plate load chokes

The EF37A has a plate resistor, I'm using grid chokes on the grid of the 2A3.

Sounds like the first thing I shoud do is check the freq resp out of the EF37A... If it's the same for both amps, then my issue is the OPT's (which we will live with), or the 2A3 cathode bypass caps.

So to measure the freq resp of the input/driver stage, I connect the scope across the 2A3 grid choke. Is that the right way to do this?

 

[kevinkr]

The EF37A has a plate resistor, I'm using grid chokes on the grid of the 2A3.

Sounds like the first thing I shoud do is check the freq resp out of the EF37A... If it's the same for both amps, then my issue is the OPT's (which we will live with), or the 2A3 cathode bypass caps.

So to measure the freq resp of the input/driver stage, I connect the scope across the 2A3 grid choke. Is that the right way to do this?

Choke issues may still apply even with my misinterpretation should one present more capacitance to the plate circuit of the EF37A than the other...

In terms of measurement that is one way, but will only tell you whether there is an error at that point or not - if there is you will need to sub resistors for the chokes and remeasure..

 

[oldeurope]

...Both amps are about the same at the low end. The scope looks good down to 10hz. That seems to be as low as I can measure with a 60Mhz scope?...

See page 6-3 Bild 2

Kind regards,
Darius

 

[sgerus]

Everyone, Thanks for the help so far, here’s an update.

I tested 5 different pentodes on the Left amp… high end freq response was about the same each time, about 35KHz@-3Db

Swaped the 2A3’s, same story.

Replaced the existing 2A3 cathode bypass R&C with a 100uf oil cap and a new resistor.
Same story.

Measured the frequency response at the grid choke of the 2A3, it’s about 105Khz@-3Db
(that’s the same on both amps)

I did notice that the frequency response makes a dip then comes back up.

32KHz = -2.8Db
42KHz = -7.0Db
60KHz = -3.0Db

The right amp makes a similar dip but it starts at 58KHz.

Tonight I will check the output of the pentode, stepping through the 30KHz to 50KHz range to see if that dip in freq response is caused be the input / driver stage.

If that test's out BAD, I will replace the grid choke with a resistor and re-test.

Right now it looks like the issue is with the OPT’s

 

[kevinkr]

Everyone, Thanks for the help so far, here’s an update.

I tested 5 different pentodes on the Left amp… high end freq response was about the same each time, about 35KHz@-3Db

Swaped the 2A3’s, same story.

Replaced the existing 2A3 cathode bypass R&C with a 100uf oil cap and a new resistor.
Same story.

Measured the frequency response at the grid choke of the 2A3, it’s about 105Khz@-3Db
(that’s the same on both amps)

I did notice that the frequency response makes a dip then comes back up.

32KHz = -2.8Db
42KHz = -7.0Db
60KHz = -3.0Db

The right amp makes a similar dip but it starts at 58KHz.

Tonight I will check the output of the pentode, stepping through the 30KHz to 50KHz range to see if that dip in freq response is caused be the input / driver stage.

If that test's out BAD, I will replace the grid choke with a resistor and re-test.

Right now it looks like the issue is with the OPT’s

Hi Scott,
You don't need to do any further testing, whether or not you realize it you have exonerated the input stages, grid chokes, and cathode bypass caps.

Your measurements show that both amplifiers test out identically to 105kHz -3dB at the chokes - therefore the drivers and chokes are eliminated as the cause.

You subbed another cathode bypass cap with no change.

The measurements you then cite are evidence of an ultrasonic resonance in the output transformer resulting in the large dip in frequency response.

You should probably contact the vendor and see if replacement is possible, if not it is outside of the audible passband and its effects are hopefully not audible. It would be nice if they both matched - obviously there is some difference in construction - in some instances reversing the primary and secondary connections (to maintain the same phase relationship) will help.

 

[sgerus]

in some instances reversing the primary and secondary connections (to maintain the same phase relationship) will help.

Thanks Kevin, I'll look into that tonight.

There is a chance that I might then the 8R output wires from the OPT wired differently on the Left amp!!!!

Also, thanks for that info on your 6j5/6sn7 srpp input driver stage.

I'm doing input/driver testing now... planing for amp#4

 

[sgerus]

I just looked at some photos of the left and right amp. I have them both wired the same but I have the 8R output wires reversed. The orange wire is the one that’s at the bottom of the 8R output (looking at the schematic). I have this wire as my positive output and the other wire as the negative, which is grounded. I have the output of both amps out of phase!!!!

Will see if this makes a difference!

 

[kevinkr]

I just looked at some photos of the left and right amp. I have them both wired the same but I have the 8R output wires reversed. The orange wire is the one that’s at the bottom of the 8R output (looking at the schematic). I have this wire as my positive output and the other wire as the negative, which is grounded. I have the output of both amps out of phase!!!!

Will see if this makes a difference!

Given the differences in interwinding capacitive coupling from one end of the winding to the other it is more than likely this will make a big difference in HF response, and reduce the measured differences to insignificant. Obviously use the amp with the better HF response as your guide. Having the outputs in phase will also make a huge difference to the way it sounds as well..