• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

How does class AB2 sound?

how does class AB2 sound?

Hi y'all

I would be very interested to know the difference in sound between AB1 and AB2 output tube configurations.

Is it better for Hifi than AB1, does it sound better for guitar use than AB1?

I know you can get a lot of power from it provided you have a stiff enough driver stage with low Zo ( i would use MOSFET followers)

I am at a turning point with a guitar amp design where i am deciding.

Cheers
Craig
 

KSTR

Member
Paid Member
2007-07-17 2:35 am
Central Berlin, Germany
Sch3mat1c said:
I doubt it matters to a guitar amp. In the name of harmonics, Eh?
IHMO it matters a lot with MI amps. With cap coupling it gives that "fart out" effect with hard overdrive, because the positive grid currents de-bias the output stage and the caps can only slowly charge back to the nominal voltage via the biasing network. Also, with a weak driver, it gives additional distortion when the driver cannot handle the pos. grid current. E.g., say, 4xEL34 driven with a 12AX7 cathodyne p/s, when overdriven it gives nice asymmetric distortion and a certain "thump".

When AB2 is used one must keep a close look at grid dissipation and runaway issues....

- Klaus
 
Craig405 said:
So generally AB2 is a good scheme, why does it generate more harmonics? are these even or more pronounced odd orders?

When there is no grid current, there isn't any voltage drop across the Zo of the driver. When current flows, that Zo causes a voltage drop. Therefore, the positive half cycles have a lower Vp than the negative half cycles, and so receive less amplification. This is also why a Lo-Z driver is so important, and why a MOSFET source follower is such a good grid driver: it has a much lower Zo than any hollow state device.


KSTR i assume you mean runaway with the solid state drivers? i wasnt aware that tubes suffered from current runaway.

Well, they do. This wouldn't happen if all tubes were filled with vacuum, but, of course, they aren't. Ionization effects of the residual gas can lead to runaway.
 

KSTR

Member
Paid Member
2007-07-17 2:35 am
Central Berlin, Germany
Craig405 said:
KSTR i assume you mean runaway with the solid state drivers? i wasnt aware that tubes suffered from current runaway.
Grid current is only one of the aspects here, in guitar amps. I've experienced runaway (not so severe as to lead to self desctrucion, but still noticable) also with fixed bias output stages which happened to run on a little low heater voltages... when driven hard, overall temperature increases, increasing emmission (and currents). The problem really starts when the PSU xformer saturates with excess current draw, then usually the bias drops, leading to even more current draw -- with often results in burn-out (same type we see with short power-downs of a few seconds, many amps are not failsafe in this regard).

In AB2, with low impedance drivers which can also sink enough current (by use of fairly low cathode/source resistors or adequate current sources) the risk of grid runaway is not very high. Dissipation is a little more problematic, but gold grids can handle a quite a bit of power, though. Also, sometimes the grid wires change the correct alignment when they expand under thermal stress, which can lead to unstable behaviour of the tube characteristics...

A great book on tube reliability etc:
http://www.pmillett.com/Books/intro_Tomer_1960_Getting_the_Most_Out_of_Vacuum_Tubes.pdf
from Mr.Millets extremely well sorted online library, a favourite of mine among the online tube related resources.

- Klaus
 
Miles how does the plate voltage get lowered?
im interested but i dont quite follow.

I can see how the current (normally zero) supplied by the driver stage will begin to develop a voltage accross any z's in the circuit and therefore the voltage at the grid of the 807 will be divided by some ratio.
Forgive my basic knowledge im still learning a lot, this project was for me to have a go at design.

Thanks for the link to pmillet KSTR ill have a look for reading material!

Also what are good MOSFETS to consider for the driver? i read in a thread somewhere the driver benefits from being run at a lower supply voltage thn the 807's plate voltage, i think to reduce dissipation when current is being drawn by the grid. So the drivers will not need to withstand the 500ish volts im going to give to the 807's.


Cheers
Craig
 
Craig405 said:
Miles how does the plate voltage get lowered?
im interested but i dont quite follow.

I can see how the current (normally zero) supplied by the driver stage will begin to develop a voltage accross any z's in the circuit and therefore the voltage at the grid of the 807 will be divided by some ratio.

Yes.

Also what are good MOSFETS to consider for the driver? i read in a thread somewhere the driver benefits from being run at a lower supply voltage thn the 807's plate voltage, i think to reduce dissipation when current is being drawn by the grid. So the drivers will not need to withstand the 500ish volts im going to give to the 807's.

The driver doesn't require the same voltage as the 807 plates since it doesn't have to swing the same voltage. If your input swing is 100Vp-p, then the MOSFET negative rail voltage need be not much higher than -100Vdc, and the positive rail can be about +40Vdc, if the grid bias is the stated -29Vdc. That's another benefit. The MOSFET can swing near rail-to-rail, though you shouldn't do that since the device internal capacitances go wonkey when the Vds becomes very small. Best to give it some headroom to keep those capacitances more linear. That way, the Source Follower will remain sonically benign.

As I stated in the earlier post, the Source Follower has a much lower Zo than any cathode follower can provide, and it avoids the complications of xfmr coupling to the finals. Also, due to the MOSFET's low r(don), there won't be any problems in sourcing whatever current the 807 grids will demand (ain't much: the 807 isn't a hard load).
 
Thanks, great info, i never considered running the mosfets at rail voltages around the swing they will need to provide.

I was still thinking inside the tube box where you still need supply headroom to get the current through the Rp. Theres plenty of 200/300v mosfets out there i bet.

You know these tubes too well! i was indeed going to run them at -29v..

Regards
Craig
 
Craig405 said:
I was still thinking inside the tube box where you still need supply headroom to get the current through the Rp. Theres plenty of 200/300v mosfets out there i bet.

Yup, lots. There are also higher voltage versions as well. One thing to look for is the lowest reverse transfer capacitance, since this capacitance becomes your Ci when used as a source follower.

You know these tubes too well! i was indeed going to run them at -29v..

I used 807s in a project, though I kept them in ClassAB1 since I didn't really need that much power. Class AB1 is good for 30W, and that's sufficient.
 
Hi, Resurrecting my AB2 thread here..

I have progressed a bit with this project i have finished the design and have a couple of questions.
Its a pretty standard design i think, i havent gone way out there, so im confident it works:

2x anode follower 12ax7 stages with 120k anode loads and 1k5 cathode resistors, the usual LTP using 12ax7's with 100K and 82K on the anodes followed by some MOSFET followers to drive the 807 grids.
I have just noticed that almost every 100w design (mine should be about 75W) uses 3 12ax7's i.e. a couple of extra gain stages.

I calculated my gain to be sufficient to produce easily enough to exceed the 100v swing required by the 807's grid (assuming 10mv peak guitar signal)..
will i be ok? or will i need an extra stage due to somthing i cant se but amp manufacturers can?

Question 2:
are the Hammond 1650R or 1650N good choices for OPT? i need 75w power with 4k-5k load.
http://www.hammondmfg.com/1608.htm

Cheers!
 

kd0bdk

Member
2008-03-14 2:10 pm
wether class a, ab1, ab2, b1, or b2, inverse feedback cuts into your gain, but does wonders for sound quality and for output impedence...for sound quality you might check eico's hf20...for a unique approach, try the macintosh mc-60...the end plate says something to the effect of Caution:while this amplifier is rated 60 watts continuous output, transient signals may reach outputs in excess of 120 watts!...don't touch the "70"volt line out or you'll have the music in you allright, and you won't even need the speakers!
 

TonyTecson

diyAudio Moderator Emeritus
2003-05-29 2:57 am
Maybunga, Pasig City
I would be very interested to know the difference in sound between AB1 and AB2 output tube configurations.

the way i understood it is that ab1 and ab2 differs only in that ab2 allows g1 to go positive instead of just negative in ab1...

and as long as your input signal does not make the output tube grid g1 traverse into the positive region, then those two should be the same, did i get it right?:D