Interstage PI fed PA project

[stephenl]

IMG_2036.JPG

IMG_2035.JPG

I finally had some time to finish this up, mostly. It still needs knobs, headshell, and a little tube rolling/tweaking.

I lifted the idea from a Hifi design on DIY Audio as I previously posted. The PA also uses Power Scaling.

The preamp is Fender Clean / JCM800ish Lead - shunt channel switching with JFETs. Tube buffered parallel effects loop. Tube reverb with 12AT7 push-pull driver (120kohm:8ohm reverb tranformer) and 12AX7 recovery and mixer. These circuits were all implemented using LP circuit boards that I've used previously - I like how they sound/function. I also wanted to be able to compare the new amp to one I had previously built with the only difference being the PA. My guitar amp interests lately have been more on the PA end.

The power-up went fairly smoothly - just a few backwards pot connections...
So far I'm really happy with the sonic results. I wasn't really sure what to expect from the PA design since it was unproven. It's always hard to describe how an amp sounds (subjective adjectives) but it covers the full bandwidth well - no deficiencies in bass (no flub), mid, or treble. Note separation / articulation is very good. Great sustain. Very Smooth.

I will definitely use this PA design in future builds. To me, it sounds better than anything else I've ever played through

The other thing I tried is partial cathode biasing. The unbypassed, shared cathode resistor is about 15% of what the normal cathode resistor would be - adding a little feedback. This was a suggestion from Gingertube (thank-you).

If anyone is interested in trying this, PM me for the Heyboer IT part #

[martin manning]

Interesting design... lots going on there!. What is the purpose of the 300k's from anode to cathode on the KT88's, nd can you explain the purpose of the 1k resistor between test points 1 and 2 and the cathodes? What is LP circuit board? And finally, how does the reverb sound, and can you identify the 120k p-p driver transformer?

[stephenl]

Interesting design... lots going on there!.
I was worried I had taken on way too much

What is the purpose of the 300k's from anode to cathode on the KT88's
I actually ended up using (2) 150k, 3W resistors. It's not my "trick", but my understanding is that it helps snub transients due to inductive kickback from the speaker when the PA is driven hard, also if the speaker is disconnected it provides some loading. I'd like to hear your thoughts on that...

, and can you explain the purpose of the 1k resistor between test points 1 and 2 and the cathodes?
Instead of the 1R resistor scaling 1mA = 1mV, it's just scaled by a factor of 10

What is LP circuit board?
L0nd0n P0wer / K0C


And finally, how does the reverb sound, and can you identify the 120k p-p driver transformer?
The reverb is awesome. The combination of the push pull driver and the cap multiplier in the main power supply keeps it quiet and clean. The cap multiplier does a great job of reducing ripple, but doesn't make the supply too stiff. The reverb driver xformer is one that I purchased from K0C - his design but manufactured by Hammond. It's a potted, PCB mount design I believe it is 125K:8R. I used an accutronics 3-spring long reverb. The circuit uses a 12AT7 for the PP driver and a 12AX7 for the return and virtual earth mixer. It has the usual level and dwell controls.

[stephenl]

PREAMP.jpg

[martin manning]

I guess the snubber R's will take some edge off of a big spike. They are bleeding a couple of mA through your 10R current sense resistors, upsetting the cathode current measurement and creating some positive feedback. Maybe you should just connect them to ground? I was asking about the 1k's connecting the test points to the cathodes. Those don't seem to be doing anything.

[stephenl]

The 1K/10R resistor combination for the testpoints is a K0C thing as well. I asked him the same question years ago, no answer. I just blindly carried them forward. I normally used the usual 1R for the current sense r's - but I wasn't sure if there was a reason for 10R in conjunction with the 300K A-K snubber resistance... It all worked fine so I just re-used it.

[martin manning]

I would guess the 1k is to prevent oscillation when connecting a meter to the test points, but I have not heard of any such problems. 1R vs. 10R is not much different as far as the current sensing is concerned, except that you do get another significant figure with 10R. In your case it is a significant part of your Rk. The 10R is insignificant to the 300k a-k snubber, but connected the way it is does introduce an error into the indicated cathode current.