12B4 pre output cap and LF instability

[pjanda1]

I'm getting closer to getting my amps working. (12B4 pre and an RH84 built into Akai M8 chassis). I've got them quiet enough that I can now hear low level motorboating (sounds like a heartbeat) from the 12B4 stage with the output cap I originally intended to use (5uf) when driving just the RH84 (30K load). When I use a 1.5uf cap it seems fine. When I step up to 2uf, I can hear some pulsating. I want the pre to be stable driving either just the RH84 or driving it and another 20K or so in parallel. When I try a 12K load, it is fine with the 2uf. I haven't tried 5uf into 12K.

So, is the worst case corner frequency of 9hz (1.5uf into 12K) low enough? I ultimately plan to insert a sub that will have significant output down to 19hz or so. Do I make the output cap switchable (once value for driving just the 30K load, another value for when I'm also driving something with the pre out). Or, do I have another problem to address? (Each 12B4 has it's own 33uf cap isolated from the rest of the rail by a diode and a 2K resistor).

Paul
Wild Burro Audio Labs - DIY Full Range Speakers

[Minussss3db]

love the Burro yes....
(and in the 3rd hand ...A burro with a white star on its forehead) peace to you....

so 2uf in 30k motors....and 2uf in 12k does not....so this tells something.

time constant for non resonant behavior??? If you ask me 9hz is fine (1.5uf into 12k, and fine at 30k...) .....So how does it sound...does it have enough grunt in the low end?

OK motorboating can also be an output to input thing...so look also at grid resistance i.e. maybe 250k, instead of 100k...

and possibly the 33uf is too big....sorry of course or too small....try a 20uf, and a 50 uf. electrolytics which are cheap just to find out....

-3db

[Zen Mod]

you need better isolation for 12B4 stage

assuming that Iq is in range of 15mA , that's certainly critical ; besides that - 2K33uF combo have corner frequency somewhat too high for my liking

solution - more uFs for decoupling ........ or reg .

choose your poison

[Eli Duttman]

PJ,

You have a "check valve" decoupling network in place, now. IIRC, you are resistively loading the 12B4. Switch to 10M45S CCS loading to improve PSRR and linearity. Eliminate a 12B4 cathode resistor bypass cap, should 1 be present. ZM is correct about reducing section to section interaction. Along with regulation and increased decoupling capacitance, CCS loading provides the necessary isolation.

The high pass pole between preamp and power amp sections should exhibit a F₃ <= 5 Hz.

[pjanda1]

Thanks guys.

Boy, I thought 33uf (per side, mind you) would be overkill. How big should I go? My B+ ended up higher than PSUD guessed (which is good for the RH84). I wasn't initially anticipating needing a 2k resistor (assumed more like 1K). I'm not using a cathode bypass now. I figure I'll have more gain than I need no matter what. I wanted to give resistor loading a shot, but it sounds as if I'll be ordering caps anyway, so I may as well get the chip.

Paul
Wild Burro Audio Labs - DIY Full Range Speakers

[Eli Duttman]

Pad the preamp I/P down, to dispose of excess gain.

If you CCS load the 12B4, roughly 30 V. are needed by the 10M45S. You can increase the decoupling filter resistor and eschew increasing the decoupling cap. value. 33 μF. should hold enough energy.

[pjanda1]

Eli,

In the past, you haven't been a fan of the 10M45S with the 12B4. It'd cost a couple of bucks more, but is something like the EC CCS a better option? I think I could get cascaded fets in there.

Paul
Wild Burro Audio Labs - DIY Full Range Speakers

[Eli Duttman]

In my 12B4 line stage, which can drive the fugly 10 KOhm IHF "standard" load, I want a high quality CCS as the load to approach infinity as its dynamic impedance. IIRC, you are asking the 12B4 to drive something much less challenging. A mediocre CCS should be good enough here.

[pjanda1]

The load will be in the neighborhood of 12 KOhm: the 30K RH84 (pot on hand) in parallel with the 20K digital signal processor driving the woofers and subs. That tough load is why I'm building the pre and not just the RH84 circuit. I built a chip buffer for the DSP (and left it at 20K), otherwise the combined load would be around 5 KOhm (due to the 6.2K load of the DSP on its own)!

Paul
Wild Burro Audio Labs - DIY Full Range Speakers

[pjanda1]

I should add that my plan is to only run the RH84 above 150-200hz. I'll put a switchable .033uf cap at the front of the RH84 to create a high pass. I figure this is a good use of the smallish stock OPT. The preouts of the 12B4, on the other hand, need to go as low as possible.

Paul
Wild Burro Audio Labs - DIY Full Range Speakers