Even order harmonics versus tube matching

5150loveeddie · Post by **5150loveeddie** » Fri Nov 25, 2005 11:09 pm

We all know that tube amps produce "even order harmonics", waveform is assymetrical making a smooth warm sound, not harch, lifeless sounding odd order harmonics that solid state circuit will give (symetrical waveform...). My question is what will increase the "even order harmonics" even more? Would some mismatching (a little...) of tubes will increase them? What other things will?

novosibir · Post by **novosibir** » Sat Nov 26, 2005 10:01 pm

Would some mismatching (a little...) of tubes will increase them? What other things will?

A mismatch by using unmatched or different strong output tubes - speak: DC mismatch - produces extra K2 (even order), but it is the wrong way, because it drives the OT assymetrical and earlier in saturation and ends up in a lack of bass response.

The "right way" is to produce an AC mismatch to the output tubes, to obtain more even order K2! How to obtain?

If you'd use a driver tube with even transconductance of both systems in Marshall amps with the usual tail resistors of 470/10K/5K(4.7K), the (theoretical) correct plate resistor values wouldt be either 82K/94.1K or 87.1K/100K, to obtain symmetrical driving of the power tubes. But that wouldt sound clean and sterile!

With the 82K/100K there's already a slightly mismatch (because tighter R's values weren't available in these days), which produces a tad of K2.

So go and increase this mismatch by measuring your (10%) resistors and look for ones with slightly higher resistance than 100K and/or slightly lower resistance than 82K and swap them in!

But be aware, there's a point of mismatch, where the amp begins to sound muddy!

Larry (the German)

Billy Batz · Post by **Billy Batz** » Sat Nov 26, 2005 11:02 pm

Very interesting.

5150loveeddie · Post by **5150loveeddie** » Sun Nov 27, 2005 12:36 am

Tail resistors???? So that would be the two resistors between the output caps? (82K & 100K)? And what would be a ok swap value? +/- 1k? or more/less?

Thanks Larry very interesting indeed, welcome to the board BTW!!!!

novosibir · Post by **novosibir** » Sun Nov 27, 2005 8:43 am

The "tail resistors" are the row of the common cathode resistors of the (Schmitt splitter) driver stage. In Marshalls usual 470 ohms, then 10K and a 4.7K or a 5K pot to ground - and with this constellation (different to i.e. Fender, Boogie) the (theoretical) correct plate resistor values wouldt be either 82K/94.1K or 87.1K/100K

And what would be a ok swap value? +/- 1k? or more/less?

Every amp behaves different and this also belongs to your more or less symmetrical working driver tube. But mostly I'm starting with both plate resistors at 100K and then I'm going to "reduce" the first one's value step by step by paralleling a 4.7meg, a 3.3meg or a 2.2meg resistor to it.

Check it out! Much success!

Larry (the German)

Billy Batz · Post by **Billy Batz** » Sun Nov 27, 2005 8:45 am

I guess that would be the reason some amps come with a trim pot between those two resistors with the B+ on the wiper.

novosibir · Post by **novosibir** » Sun Nov 27, 2005 8:54 am

Right! But the idea behind the trim pot was to make the driver stage adjustable, to obtain its working as symmetrical as possible, because the otherwise needed resistor's values were hardly to get.

If your amp has such a trim pot, then you can use it now to adjust more or less mismatch to its AC output, to obtain the loved extra even order K2.

Larry (the German)

Billy Batz · Post by **Billy Batz** » Sun Nov 27, 2005 9:29 am

Hey welcome aboard man! Its good to have a long time tech around for once. George is great but he has a business to run and Id rather keep getting my shipments fast as hell

flemingmras · Post by **flemingmras** » Sun Nov 27, 2005 10:04 am

Billy Batz wrote:George is great but he has a business to run and Id rather keep getting my shipments fast as hell

Yeah...that's if UPS doesn't decide to fuck around for 4-5 days and loose track of your package...

Jon

Billy Batz · Post by **Billy Batz** » Sun Nov 27, 2005 10:38 am

Fact of life...

novosibir · Post by **novosibir** » Sun Nov 27, 2005 4:15 pm

5150loveeddie wrote:Thanks Larry very interesting indeed, welcome to the board BTW!!!!

Billy Batz wrote:Hey welcome aboard man! Its good to have a long time tech around for once.

Thanks guys for the kind entrance and for the flowers

I'm also happy, that I've found a board with great guys, who are delighted in building their own fuckin' great sounding tube amps.

And I have to thank George once more for that, that he'd helped me to get in touch with John from SoZo about a year ago! Because w/o SoZo and w/o Mercury Magnetics my amps wouldn't be that today, what they now are.

So also I'd be glad, if I couldt help you from time to time with my experience and knowledge in building tube amps.

So long,

Larry G

5150loveeddie · Post by **5150loveeddie** » Sun Nov 27, 2005 4:59 pm

A guy like you around is a fantastic addition for all of us, stick around often!!!!

5150loveeddie · Post by **5150loveeddie** » Fri Jan 20, 2006 10:06 pm

Larry or anyone, here is more info on even order harmonics I found...

What do you guys think?

You could just put another resistor across the tail resistor. That'd give you a bit mo' gain too. The longer the tail the better the balance and vise versa. If you got enough gain going on in you preamp you can change the cathode resistor on stage right before the cathode follower. That will make the preamp clip assymetrically. Just messing with the bias on you power tubes will get you different hamonics as well, and you don't need to be clipping for it to happen.

What is THE tail resistor? And what value should we put accross it to add K2s??

The resistor right before the the cathode follower, the 820ohms right? So we change it with what? 1.1k, 1.5k????

Billy Batz · Post by **Billy Batz** » Fri Jan 20, 2006 10:12 pm

Its the 10k in the PI. I guess you would start with something like 150k and work your way lower? I was always under the impression lowering that 10k made things really bright.

5150loveeddie · Post by **5150loveeddie** » Fri Jan 20, 2006 10:51 pm

Billy Batz wrote:Its the 10k in the PI. I guess you would start with something like 150k and work your way lower? I was always under the impression lowering that 10k made things really bright.

I really don't know Dan, will have to monkee things around and see...Maybe Larry could jump in here....

Might try the driver's plates resistor trick first....