2nd Harmonic, even harmonics

is there a general rule for increasing even harmonics in tube pre design.

or cancelling out odd harmonics...
(a lot info on cancelling even harmonics some reason but not on cancelling odd harmonics)

i found some posts and notes but not much info on this subject...

High enough signal level and small enough Rl and 2nd harmonic will be very high. There will be higher harmonics too but 2nd will dominate.



With pentodes negative voltage on suppressor grid can increase 2nd harmonic distortion and reduce gain. Negative voltage on G3 will need to be fairly high unless you use dual control pentodes.

As a rule, harmonic distortion in triodes is proportional to output voltage level. It is also proportional to load. The basic shape of the transfer function means that initially nearly all 2nd harmonic is produced but as you traverse more of the curve you get higher order harmonics. it all goes wrong when you either reach clipping or the onset of grid current. For low mu tubes like the 6SN7 and 12AU7, grid current tends to occur first unless your bias is way out or you have a very low plate voltage.

Cheers

Ian

sorry, i was breadboarding, 
now i can dial in second harmonics way better then ever
thank you Heikki and Ian, and also PRR(yesterdays posts for HPF solution)

In amplification devices like a tube or a transistor, the gain varies a little with the current going through it. So if you have a very small amount of current going trough the device, the gain is noticeably less than if you have more current going trough it. So when the tube is near cutoff the gain is less. When the input is high and the tube is very much "on", the current through the device is at it's maximum and the gain is higher. So the affect is asymmetric gain. This is the dominant source of the 2nd harmonic. So, in theory if you make the load resistor bigger, you should get more asymmetric gain because there will be a larger voltage voltage swing and thus a larger difference in current and thus a larger difference in gain and thus a larger 2nd harmonic (assuming the cathode resistor is fully bypassed). So instead of the usual 100k, you might try 220k (might require a voltage increase). But in practice I have no idea if that would actually increase 2nd harmonic significantly if at all. I don't think I would even trust a software simulation. You would need to just try it and see what it looks like in an audio analyzer.

sounds good, thank you... i rather try in real life, then simulate it 

2nd harmonic is mostly proportional to Current Swing (what squarewave said). Drive the device down to zero current and up to over twice idle current, 2nd will be large.

For most tubes, 5% or so max. This is not large on the ear.

For BJT with *no* negative feedback, to abut 26% 2nd harmonic. Interestingly nobody does this.

There's no good way to cancel 3rd harmonic. The main cause is gain droop at maximum swing, if not device droop then by running into the power rail (clipping).

thank you PRR,
very interesting re BJT, i should look in to it for sure...

Look at your load line and operating point. The more asymmetric the line spacing either side of the operating point, the more even harmonic distortion you will get. Some tubes are more useful than others if you are deliberately trying to increase 2HD.

Matt Nolan said:

Look at your load line and operating point. The more asymmetric the line spacing either side of the operating point, the more even harmonic distortion you will get. Some tubes are more useful than others if you are deliberately trying to increase 2HD.

Click to expand...

thanks Matt,
i dont have too many different tubes in my stash, but so far 5751 seems a good candidate.
6SN7 is nice too

change a pentode to triode operation and harmonic distortion usually goes down

change from single ended circuit to push-pull (see Pultec eqp1a) and harmonic distortion usually goes down

change a coupling cap to an innerstage transformer and i do not know what would happen except more people would buy it because it contains more trannsformers,

and remember, a bucket of chicken does not feed a family, it conntains a family,  ;D

Simple BJT biased and driven for high 2nd harmonic:

To get the effect being described here, i.e. asymmetric operation you should not use tubes with nice parallel load lines like 6SN7, 6N7 and ECC88 types.  What you need are semi-remote types like 12AU7, ECC81, 6BC8, 6ES8 etc.  There are many more remote pentodes that can be triode wired, raid an old radio!

Best
DaveP

PRR said:

Simple BJT biased and driven for high 2nd harmonic:

Click to expand...

thank you PRR, this is looking fantastic 
i need to order some parts this week, definitely adding couple of  BJT's to my list...
i report back how it goes on my end.

i couldnt play much with the load, but lowering  bias resistor way down was helping a lot on 5751 pre for 2nd harmonic, but nothing like this BJT.

DaveP said:

To get the effect being described here, i.e. asymmetric operation you should not use tubes with nice parallel load lines like 6SN7, 6N7 and ECC88 types.  What you need are semi-remote types like 12AU7, ECC81, 6BC8, 6ES8 etc.  There are many more remote pentodes that can be triode wired, raid an old radio!

Best
DaveP

Click to expand...

aha! i think i have some 12AU7's in my stash... thanks Dave, i will give it a try.

Above, 33mV pk input gave 28% THD out.

Theory for single simple devices predicts the THD% will scale about as the signal amplitude (below clipping).

SPICE says 3.3mV in gives 2.9% THD, 0.33mV gives 0.3% THD.

As 2.9% 2nd harmonic is hardly audible, we see that "color" happens only for the top 20dB of the dynamic range, mostly the top 10dB.

3rd harmonic theory (simple devices below clipping) says 3rd harmonic rises even faster than 2nd. Or conversely, falls-off faster at low level. However 3rd is more obvious (for several reasons, including rapid rise). Again only the top 20dB of dynamic range will show audible color.

this is where i am at now with 5751 tube

-22.5dbFS = 0dbu

lo gain


edit: as long as i keep my input level at 0dbu. i think these are looking good
looks like i am gonna need some sort of metering at the input for 0dbu, so i can mark my gain knob for
harmonics.
i can always go with ear, but VU's will look nice

5751 med gain

-22.5dbFS = 0dbu

rough schematic, but i think its pretty much accurate to my breadboard

(R1 500 ohm on input is for the passive  EQ )

in --> 600:600 trx -->passive EQ-->5751 gain--->10k:600 out

edit:
output pot is 50k for now(thats what i have handy), i will fiddle with that too... 
may be fixed resistor and move output_trim after 10k:600 trx!

PRR said:

Simple BJT biased and driven for high 2nd harmonic:

Click to expand...

Was going to post this too but PRR beat me to it 

I'll just point out that

• below about 10% THD, the scaling PRR mentions is almost perfect.  1mV at input gives 1% THD etc
• V2 has to be a true, low Z voltage source.
• If not, THD will be reduced in proportion to the source Z compared to the input Z of the BJT

ricardo said:

PRR said:

Simple BJT biased and driven for high 2nd harmonic:

Click to expand...

Was going to post this too but PRR beat me to it 

I'll just point out that

• below about 10% THD, the scaling PRR mentions is almost perfect.  1mV at input gives 1% THD etc
• V2 has to be a true, low Z voltage source.
• If not, THD will be reduced in proportion to the source Z compared to the input Z of the BJT

Click to expand...

thanks ricardo,
what exactly do you mean by the  "true, low Z voltage source"