Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Quatro on March 06, 2013, 05:53:40 pm
-
Hey,
I built a version of the "Champ Checker" talked about here:
http://www.el34world.com/Forum/index.php?topic=11281.0 (http://www.el34world.com/Forum/index.php?topic=11281.0)
I have a solid state rectifier, a 6V6, a 6SF5 for the stage before the 6V6. Then the "checker" portion. I have two sockets. Each socket is wired to run both halves of the preamp tube simultaneously (parallel). There is a switch that dumps the grid to ground alternately for each socket. Much like the drawing of the A/B Comparitor in the link.
It works, but there is one issue. I have noticed when I physically pull one of the tubes out ( to swap in the next for the a/b test) the amp gets quite a bit louder and brighter. It's like the unused tube is dragging the other down. Any ideas why and how it can be fixed?
Do I need "mixing resistors" before the coupling caps that feed each tube? Something else?
Thanks
-
No, it's always gonna happen.
The problem I see with the various designs drawn out is that each of the triodes stages impact either the d.c. operating point for all the other triodes (Richard's drawing) or the apparent a.c. load on the plate of each triode (Jeff's drawing).
EDIT: Actually, if you implemented Richard's cathode switching scheme as drawn, it would mitigate any issues with his approach.
I know these were drawn with the idea that it would speed things up to plug in multiple tubes, but they sacrifice a little something on knowing exactly what 1 tube is doing on its own.
Which was why a stock Champ circuit was originally suggested (by PRR, I think). Each triode stage has its d.c. and a.c. operation set independently of the other, and you're using the tube as you would in an amp. You could measure voltage at the cathode and plate to get an indication of internal resistance and static transconductance, and measure a.c. volts at the grid and plate when injecting a test signal to directly measure gain (which won't be mu, but hopefully you already know the how/why of that).
-
use a VVR for each 12AX7. IOW, that would be 3 VVR systems, one for each pre-amp bottle. use the variance of VR to dial in B+ then you can make rough approximations of mu/gm/Ib, etc.
--pete
-
EDIT: Actually, if you implemented Richard's cathode switching scheme as drawn, it would mitigate any issues with his approach.
I tried this switching method some time ago on another project. It works fine, but the on/off is a slow fade not an instantaneous action. What i really need is a switch that can be flipped to and fro with instant action.
use a VVR for each 12AX7. IOW, that would be 3 VVR systems, one for each pre-amp bottle. use the variance of VR to dial in B+ then you can make rough approximations of mu/gm/Ib, etc.
Too complex for this test device.
-------------------
My original intent was to have a testing platform that allowed one to compare tubes on the fly. Getting a hold of Radial's head switcher and cabinet switcher was a real eye opener for my testing procedures and I hoped to have the same set up here.
Good news: basically works fine as is. Allows instant switching between tubes.
Bad news: kind of a bummer that the device is a tone robber, but I'll take what i can get.
Thanks
-
EDIT: Actually, if you implemented Richard's cathode switching scheme as drawn, it would mitigate any issues with his approach.
I tried this switching method some time ago on another project. It works fine, but the on/off is a slow fade not an instantaneous action. What i really need is a switch that can be flipped to and fro with instant action.
...
My original intent was to have a testing platform that allowed one to compare tubes on the fly.
So you're trying to use this not to determine tube operating characteristics, but for tonal comparison of 2 different tubes? If so, you really need 2 independent channels with a shared input jack, so you can flip-flop between tubes without disturbing the d.c. operation.
-
So you're trying to use this not to determine tube operating characteristics, but for tonal comparison of 2 different tubes?
Precisely. Sorry I didn't make this clear. This about testing using ears.
If so, you really need 2 independent channels with a shared input jack, so you can flip-flop between tubes without disturbing the d.c. operation.
Sounds good. How do I make my channels independent?
-
Just like any 2-channel amp: have 2 separate paths for each channel.
Maybe instead of mixing resistors, you have a SPDT switch to connect the output of only 1 at a time to the output section.
If you were clever with solid-state (and I'm not), you could use JFETs. One in series with the signal, one from the signal to ground (x2 so each channel-output has a pair). Your switching would send appropriate control signals to the JFETs to fully turn-on the series JFET and shut off the shunt JFET for the channel in use; the unused channel would get the opposite set of control signals (which turns off that channel).
The JFET idea occurred to me only because you might be tempted to switch both the inputs to each channel as well as the outputs, and run those points to a DPDT switch. But that would invite trouble (oscillation) because inputs and outputs (with a lot of gain in between) are brought next to each other. Using the JFETs allows the switched elements to remain separated (because the JFETs are brought to the circuit, not the circuit to the JFETs), with the actual switch only changing the state of the d.c. control voltages.
Anyway, maybe that's a distraction, because you can probably competently switch just the outputs to listen to one channel (1 tube) or the other.
-
Maybe instead of mixing resistors, you have a SPDT switch to connect the output of only 1 at a time to the output section.
Mixing resistors: will this get the job done? The reason i ask, having thought about the two channel idea I was thinking two separate inputs and an A/B box to switch.
SPDT: you mean a switch downstream from each tube in between them and the next stage?
-
Maybe instead of mixing resistors, you have a SPDT switch to connect the output of only 1 at a time to the output section.
Mixing resistors: will this get the job done?
Imagine a 2-stage Champ preamp as 1 channel; duplicate everything for the other channel. Where the coupling caps would normally connect to the 220kΩ grid reference for the 6V6, have a switch. Center lug goes to the 220kΩ resistor. Outer lug goes to one of the coupling caps, other outer lug goes to the other coupling cap. You get either channel 1 or channel 2.
Mixing resistors... Well, you're problem was that your outputs had parallel paths by which the signal could sneak backwards through another stage's coupling cap and plate load. If you add mixing resistors, you have this possible problem again. I suggested a switch because then you can break the connection backwards through the channel/tube not being auditioned.
The slowed settling you experienced was very likely due to the shared cathode bypass cap and resistor; in fact, if you split those up, you might solve some of your problem. The issue is that when a tube is removed or a new one plugged in, the current through the cathode resistor changes, which changes the voltage across the resistor and the bypass cap takes a finite time to charge to that new voltage.
Anyway, I'm talking about a complete topology change to what you have now, I guess. I thinking Tube 1 = gain stage 1 and 2 of the Champ; Tube 2 = gain stage 1 and 2 of the other channel (that's identical). Input jack can feed both gain stage 1 grids at the same time. Switch between the coupling cap for each channel's stage 2 and the output tube. If you get popping you just can't like with, add a 5-10MΩ resistor from the cap's connection with the switch to ground.
Anyway, this is all far afield from the original intent of the Champ Checker, which was to be a real-life tube tester that offered meaningful readings. It seems like your original incarnation had too many inter-relationships between tubes you were listening to for the purpose you had in mind.