Split Load Idea?

[goldstache]

I built up a 6SL7 tube bass DI/Preamp.  I wanted to be able to drive Line level as well as drive the input of an amp.  The thing sounds perty' righteous.  I was experimenting with lo/hi output modes via coupling caps selected from right at the the anode, and taken from a Split load.  Its a single tube 2 stage design. The output of V1a and V1b are tied to a DPDT switch selecting both stages output. 
Kinda like this:


                        DPDT SWITCH




Selection 2                        Center Lug                Selection 1


-V1A anode 1uF    >>>>            To Grid of V1B     <<<<        -V1A split load 1uF


 
-V1B anode 4u7   >>>>             To Output XFMR    <<<<       -V1B split load 4u7


Probably could have scrawled a schematic faster than how long it took to depict the above switching:)
But. I think you get the drift.

I know there are better ways to control output gain, like a simple input pad.  But I was experimenting with the AC characteristics of split load and its effects on Output and Impedance.  Especially Z when paired with a signal matching transformer on the output making up the source impedance. 

Does someone want to take the time to help me along with the math to calculate the Z and Voltages so I don't just keep plopping in parts and measuring.


The split load has a charm to it that the linear version with pad just doesn't have though my Voltages AC are similar. 

I have a 15k/600 output transformer for balanced line output and Direct off V1B's plate/split load caps run right to the Hi-Z output. 

I know a schematic would help.  I'll draw one tomorrow if I can.

Thanks!
   

[HotBluePlates]

Are both halves of the 6SL7 being operated as split-load inverters? Or did you mean both triodes are plate-loaded (with normal cathode bias resistors), and the plate loads are split into two parts?

[goldstache]

Plate loaded cathode biased.

Cap at each triodes plate and a secondary cap at the junction of split load resistors from B+ to anode.  So 2 coupling caps headed out of each anode on each triode.  DPDT switch selects cap at both plates in one setting and cap at split load junction for lower gain setting.

I think that the output impedance of a triode stage is calculated with Ra and ra in parallel from the formula:

Ra X ra/Ra + ra

Does not loading one of the taps (i.e.- when the switch is thrown and one set of caps is just floating) cause any ill states?  Switch pops like a banshee of course too!
Thanks

[HotBluePlates]

...  Switch pops like a banshee of course too! ...

Sure it does, because you're switching the coupling caps from one d.c. voltage to another (and a fairly big voltage jump at that). There's no easy way to stop that popping.

There is a "hard way" in which the switching is done by JFETs (or perhaps optoisolators) instead of a mechanical switch, but then you need solid-state switching circuitry to control those. And ideally the control circuit would fade from one output to the other instead of instant-switch. So it can be done, but unless you're adept at building such circuits (or find just the right kit, which will be challenging), this will seem hard.

... Does not loading one of the taps (i.e.- when the switch is thrown and one set of caps is just floating) cause any ill states? ...

Nope. If you connect nothing to the cap, the tube feels as though the cap is not even there.

I suppose it's possible if you have very long wiring from one of those outputs looping around the chassis and near a high-imedance, low-level stage earlier in the circuit that you could get the unused wire to create oscillation. But I think the chances of that are fairly remote.

... I think that the output impedance of a triode stage is calculated with Ra and ra in parallel from the formula:

Ra X ra/Ra + ra

The output impedance will only be relevant for the triode connected to your preamp's output. Then it matters because you'd like to have an idea of the cable capacitance being driven (so you know how much treble will be knocked off your signal), and also because the transformer primary may look like a heavy load to a typical plate output which then knocks down the gain of that triode.

Are you using the transformer you have as a plate load (i.e., transformer primary instead of a plate load resistor)? Or are you trying to use it as a matching transformer (triode has a normal plate load & coupling cap, and the transformer primary is after the coupling cap)?

I suspect a 6SL7 won't manage to drive low impedances well, and possibly not long cable runs without treble loss. What are you hearing when you try the preamp/DI with a long cable on the output?

[goldstache]

Thanks for the insight.


Regarding the quote:
"Sure it does, because you're switching the coupling caps from one d.c. voltage to another (and a fairly big voltage jump at that). There's no easy way to stop that popping.
There is a "hard way" in which the switching is done by JFETs (or perhaps optoisolators) instead of a mechanical switch, but then you need solid-state switching circuitry to control those. And ideally the control circuit would fade from one output to the other instead of instant-switch. So it can be done, but unless you're adept at building such circuits (or find just the right kit, which will be challenging), this will seem hard."


I was already looking at relay switching options.  But Fet's would be cool too.  As I've never worked with switchers other than mechanical.  Especially the PWR Supply demands and topology.


Im using a matching transformer right off the coupling caps from the triode plate.  After crunching the OutZ I am getting 7k7 in high gain mode and 4k2 in low gain mode.  It then heads to a 15K/600 ohm matching transformer.  Not exactly bridged or matched though it seems to drive both Hi-Z, -10 line off the the preamp Hi-Z output (taken from cap atthe triode plate before signal matching transformer) and it drives a mic input on a mixing console with the signal matching transformer fed XLR jack.  It sounds good.  A little bit of noise when maxed out though.  Power supply is external and away from the preamp. 


None of my cables have been over 10' as I'm just using it as a preamp so it's close the gear it interfaces with.  Perhaps cable capacitance could act as a built in Zoebel type filter 


If you have a better transformer ratio suggestion, let me know, Im trying to get the noise floor down and proper matching may give me better signal to noise ratio. 


Thanks!

[HotBluePlates]

What are the plate load resistors? 7.7kΩ & 4.2kΩ sound improbably-low for plate outputs of a 6SL7...

[goldstache]

First stage is 100k plate 2k2 cathode with a 100uF bypass.  Second triode is 10k plate 270E cathode bypassed with 470uF
315v B+.  There's a baxandall and gain control between the two stages.  The split loads are just halved.

V1a 47k 47k
V1b 4k7 4k7