The Kuartlotron - keantoken's simple error-correction superbuffer

I wanted to know if the new board will only require a rectified bipolar power supply like the one pictured?

Also, if I us a 100K stepped attenuator on the input do I need the 100K resistor?
 

Attachments

  • Rectified_Power.jpg
    Rectified_Power.jpg
    11.2 KB · Views: 885
I wanted to know if the new board will only require a rectified bipolar power supply like the one pictured?

Also, if I us a 100K stepped attenuator on the input do I need the 100K resistor?

Yes. There's no rectifiers on the board Keantoken just posted. A 9v x2 transformer will work nicely with the new regulated boards, otherwise it's a bit high and 7v would be better for the current boards that I'm using without regulators.

I would leave the 100k in, unless you want to see just how powerful your source is...
 
I just got my boards today, thanks keantoken for the quick shipping.

They are nice and small. I do have a couple of questions before I start the build.
1. These boards required that I make the 15pF 270R connection off the board?
2. The boards do not indicate the transistor types. Are they installed in accordance with the attached picture?

I'm still waiting on a couple of parts. The RJM X-reg board, ebay DACT 100k stepped attenuation, input selector switch, and a chassis.
 

Attachments

  • Kuartlotron.jpg
    Kuartlotron.jpg
    382.1 KB · Views: 834
I hope this is your sense of humor; that or you're going to send me a $3k hot air bed station....
It is either that smaller layout or the added coil for the cost of a fast thing in an overlarge size. Seems like the needless coil is cheaper at the moment.
I did mine coin size with through-hole parts, just because that's what it looked it like it needed. Well, it was triple layer and not the prettiest thing; but, I've still got to say that when it is small enough, then the coil is not needed. I was told that it was supposed to be fast, so I built it for fast. Well, you wouldn't do otherwise. But, if you did, that's when you need the extra coil.
Sure, the new SMD things are awful, at least 1/3rd off the mark; however, they're a lot better than a foul layout. And, for a fast thing, a foul layout is one that is too big.
In this application, there is some leeway to consider using SMD parts.
 
If the transistors are slower, then that could adversely affect stability. The effect of the inductor can be simulated to work without parasitics added, because the inductors is compensating for the slowness of the transistors, not for layout.

The total trace length of the PCB inductor is 48cm. If it were adjusting for layout inductances, that would be too large (several times larger than the longest trace on the board), especially since the ground plane reduces trace inductance.

In terms of capacitive parasitics, the transistors themselves still have more capacitance than the traces that connect to them, so are the limiting factor. But the inductor isn't really compensating for the transistor capacitances. It's compensating for the junction transition time. The B-E junction behaves like an inductor at RF and this can cause too much phase shift even if Cbe is very small. This is why some circuits are near impossible to stabilize when you only think in terms of Cob or Cbe.

So again, even for inductances the transistors are probably slower than the layout.

I imagine the inductor plays little purpose except for edge cases, but having it there improves reliability in some cases and in simulation even improves distortion. If anything about layout worries me, it's pcb leakage. I forget what the general value is but I thought somewhere around 500Mohm+1pF or something like that. But that is swamped even by the 100k+15pF input impedance, and most DIYers don't seem to have had a problem with it.