ampeg b15 inverter

[catalin gramada]

Hi.
Looking at b15n heritage schematic I found third part of bias voltage for this stage is developed over feedback network due 1.220k and 3.3k+OT winding resistance in parallel. Did not found any electrical explanation what for. This model is a "optimised " replica of original. Do You think the designers have some reason to compensate some in this way? Seems have more sense to move feedbak return directly to junction between 1k bias resistor and 220 ohm feedback "shunt" and not to the ground.Wiring feedback return to the ground meant nothing but troubles as time common cathode resistor is fully bypassed (if this capacitor is not a high quality one the stability of this stage is compromised btw.). I don.t get it
. Can someone explain why, please? Thanks

[catalin gramada]

More specific :

[Enzo]

It is not a long tail pair phase inverter. The NFB goes to the upper triode, the lower takes its signal off the 510k/470k divider.

[catalin gramada]

It is not a long tail pair phase inverter. The NFB goes to the upper triode, the lower takes its signal off the 510k/470k divider.

Thanks. I asked for nfb return btw.

[nickb]

Hi.
Looking at b15n heritage schematic I found third part of bias voltage for this stage is developed over feedback network due 1.220k and 3.3k+OT winding resistance in parallel. Did not found any electrical explanation what for. This model is a "optimised " replica of original. Do You think the designers have some reason to compensate some in this way? Seems have more sense to move feedbak return directly to junction between 1k bias resistor and 220 ohm feedback "shunt" and not to the ground.Wiring feedback return to the ground meant nothing but troubles as time common cathode resistor is fully bypassed (if this capacitor is not a high quality one the stability of this stage is compromised btw.). I don.t get it
. Can someone explain why, please? Thanks
[ATTACH=CONFIG]44594[/ATTACH]

You could use a bigger capacitor for the 25uF but it would have to have a large value, be bipolar and would have added cost to the design nor would it offer any advantage.

[Enzo]

I was implying that the NFB went only to the upper triode on purpose. The cathode bypass cap keeps the NFB away from the lower triode. If you moved the NFB to the cop of the cathode resistor and cap, you won't have much NFB any more, the 25uf cap will squash it.

The NFB signal comes from the 4 ohm tap of the OT, the 3k3 resistor and the 200 ohm resistor form a voltage divider as far as the NFB is concerned.

All of that is why you don;t want to move the return spot.

[catalin gramada]

So, first drawing is a better wiring arangement from a NFB perspective ?

[Enzo]

The second drawing requires the speaker outputs be isolated from chassis.

[Chuck H]

So, first drawing is a better wiring arangement from a NFB perspective ?

[ATTACH=CONFIG]44614[/ATTACH]

Stick with the first drawing. The NFB loop is an AC voltage driven circuit and the 25uf bypass cap accounts for that. This action makes the 200R resistor the shunt as shown in the first circuit diagram. Your second diagram would appear to 0V the OT secondary via the 1k cathode resistor. I wouldn't recommend it

[catalin gramada]

Hi. I don.t feel any reason to keep secondary OT grounded other than sort of safety measure in a possible but improbable scenario. On the other hand I see a reason for instability in this stage using anything but a very high quality capacitor (those capacitor is an AC shunt dependable by its value and its performance to offer a very good decoupling in high frequency range). The third reason comes from biasing consideration forcing the circuit to share bias voltage through feedback network.(above third part of bias voltage is developed in feedback network in this case).We don.t talk about a dc balance here but still restrict the biasing choices. So, do you see any incovenients in a second drawing, other than safety consideration please?

late: never heard but if grounding OT secondary is a compulsory safety standard I have to ask to delete this thread. Thx.

[catalin gramada]

I builded one for a test. Better works with individual cathodes res. Using 12AT7 it was able to drive nicely a KT88 pair. 7db global feedback injected with no stability issues (cannot inject more cause put into the clipping before power stage reach the full power) . voltage amplification 15. Pretty well balanced near before clipping 2.5% without messing arround. It is a good one.Think I.ll use it into a sort of clone

[catalin gramada]

...and suited with grounded secondary OT,by choice

[nickb]

I really feel you are missing the point of the original design. You can come up with different arrangements with the equivalent functionality but the goal, I think, was lowest cost/component count.


If you really don't like insignificant the DC current in the winding, you could just add a cap in series with the 3.3K feedback resistor.

[cjenrick]

Original B15N used the upper tap on the OPT sec. for feedack. Heritage looks like the switched it to the 4 ohm tap,

so the 3.3K was probably changed to compensate for less voltage from the OPT sec.

still, there is a ton of NFB any way you slice it, so try experimenting with higher values. This circuit tends to distort pretty easily, anything past 4 on the volume and you have a wet sponge, the 1 Meg linear pots on the original chassis do not help this cause,

what do i run? 1 Meg audio taper volume pots, 47K FB resistor, a 6SN7 inverter instead of the 6SL7, (less gain) a V4 output transformer with a pair of NOS Tungsol 6550's running at 550 B+ , and Lord have mercy i never had it so good, 15" + 4-10's with a Jazz bass,

interesting offshoot, same circuit with Bendix 6384's, four of them, only two get an input signal, the other two just drag idle current thru a cathode resistor, this was done by accident, started as a stereo amp but only one channel needed, what this does is regulate the power supply by keeping a constant load on the B+. amazing sound give it a try sometime.

[catalin gramada]

Used one 12ax7. one 12at7 and a kt88 pair got clean 100w sine before power stage clipping for flat setting. 20mv input sensitivity. I much prefered early B15-NA design with tone control after first stage and ss rectifier