The official G9 help thread

[knight84]

Thank you Bruce for this advice. I did the test. The phase switch didn't change the resistance value. Both channels have around 147-150ohm in both of the phase positions.
For the 12 VDC maybe I made a wrong alarm I think. The relays and light is working properly and I measured 11.9VDC again.
I think I will leave the problem of gain and phase for a while, then come back with a fresh mind.

 

[bruce0]

Jacob: any reason he might see low gain (6 db low)?

 

[gyraf]

The only way I can see the lower gain happening (in a correctly built unit) is by natural variation in the raw open-loop gain of the tubes. I often see 3-4dB variation in a batch of new tubes, and two low-gain ones would explain.

On the other hand, the lower-gain tubes often has much less self-noise, being good for input stage amplification..

Jakob E.

 

[Deepdark]

Hi guys

I was looking to put some Edcor on the output instead of Lundahl/OEP. I first look at the XS1100, but the pin layout isn't the same than the Lundahl, so I can't really figure out how to match it correctly. After, I was looking at some XSM10K/600/2.4K/600 or WSM10K/600. Anyway, the question is, how to wire it on the board. On those models, Pin 1 and 4 are the input + and - and pin 5 and 8 are the output + and - and there is a CT on the primary and the secondary. If I connect the input pins 1 and 4 into the Lundahl pin 2 and 5 holes and the output pins 5 and 8 into the Lundahl pins 7 and 10 is it ok? What about the center tap? where do I have to connect it? Do I have to isolate the tranny, or install it far from the pcb or there no trouble with it?

Thanks for your help

 

[Deepdark]

up!

 

[bruce0]

Search this thread for inductance, and the site for g9 inductance.  There are several g9 posts on the issue and how higher primary inductance on the output extends low end response.

 

[Deepdark]

I was looking at the schematic and the printed circuit and have a question about the input transformers. When I look at the connections for Lundahl, it seems to be wired 1:5, so 200:5K? Am I allright? But, when I looked at the OEP connection, it seems to be wired in series, so 600:25K. Is the input impedance has to be at 150/200 or 600? and the secondary impedance has to be around 5K? My guess is that the OEP shoul work in 150:6.25K. Sorry for the newby one haha. Just try to figure it out.

 

[bruce0]

I haven't checked your math, but it does sound right.  Transformer matching is "rough justice" sort of thing.

The goal is to bring the impedance of the inputs and the output of each interface close together, resulting in the best power transfer (Wiki impedance matching).

In the transistor world, it is less of an issue in audio equipment, as the process with transistors has allowed voltage "bridging" (terrible name, there is probably a better one) to be used.  This is just the process of making the output impedance much less (traditionally 1/10th) the input impedance on each interface.  In summary the receiver is always high impedance (think resistance) and thus a voltage drop develops across the receiver inputs and that is amplified, recreating the power in the signal in each step.

In the transformer your goal is to move POWER (I. E. watts, amps, VA etc) down the signal chain, because tubes (as typically used in audio) have less ability to produce that power than transistors.

There are lots of mics that still have output transformers, but they are always the first step in the chain so they can be connected to either low noise transistor inputs or transformer inputs. 

Anyway, the best power transfer comes from perfectly matched impedances, but since the impedance changes based upon how you terminate the transformer, and what the impedance of your mic (in this case is), it is a big rough justice balancing act.

 

[Deepdark]

thanks a lot! Maybe the difference between the two ratio / impedance of the lundahl/oep explain why one run the low end better than the other?? I was planing using a Cinemag CM-75101A, wired 1:10 (150:15K) in input. Does it looks allright? Should I wire it 1:5 (600:15K)??

 

[nootje]

building a channelstrip of the G9 and I came up to this:


R11 amd R12 aren’t connected as in the original schematic. Cap C7 (molex pin 5) is connected to R11 and R12.
The other side of R11 is connected to the switch and R13 that trace should not touch R12

On the PCB i colored the traces. I can see pin 5 is only connected to R11.
R12 is connected to the other side of R11 and connected to R13 and the gain switch.

In the schematics (with the colors) I get this: red to R12 is cutted and orange to R12 is created

Is this on purpose or is this an error? what might be the influence on the signal? only different Gain settings? I’m trying to learn on this

Greetings,

Chris

 

[bruce0]

Looking at the traces, you are correct, that junction is open BUT it is connected by R11 which is only a 1K addition to the 1M resistor.  (0.1% increase in value for R12).  You have drawn this as an orange dotted line.

So yes the schematic is drawn with an error, but the function of the circuit should be unaffected as the net result is inconsequential additional current through R11 and inconsequential additional resistance on R12.

As you point out the junction of R13 and R11 is connected via a short trace to R12.

I think it is really great that you found this error.  Have no fear the error will not change the circuit.  But I am sure Jacob will want to update his schematic.

How did you find it anyway?  Are you relaying out the board for a channel strip.

 

[nootje]

well, the chanelstrip involves a single G9, a CTQ and a LA4A style compressor.

For the single channel G9 (lets call it a G4.5 ;D ) I chopped up the whole pcb into different stages. 2 reasons, first of all, there is no longer need for the low cut, since i have one on the CTQ. Another reason is the second channel only absorbing a lot of space in the case.

The pcb limitations in that free eagle version were good enough to make an psu, input, output, gain and 48V stage. I also added an extra led VU.

Every sub pcb is recreated in eagle starting with the schematic and then retrace the whole thing. I recheck my work with the existing pcb and ...boom, there i saw that small error 

a preview of the chanelstrip frontplate?

Chris

 

[Deepdark]

Hey guys, anyone here who have a tricks to insert a tube socket on a ocb? Damn there hard to plug into the pcb

 

[audiophreak]

I usually get one or two in, then with a small screwdriver work each leg individually , hope that helps 

 

[Deepdark]

yeah I tried this way but the holes are damn small for the legs. I plan maybe squeeze the legs with some pliers, or just try to enlarge the holes.

 

[Gustav]

yeah I tried this way but the holes are damn small for the legs. I plan maybe squeeze the legs with some pliers, or just try to enlarge the holes.

The sockets I use are a perfect fit, so they are out there. Check the diameter of the legs when ordering.

For a socket with legs too wide, I would cut them to size with a plier.

Gustav

 

[Deepdark]

Thanks Gustav. That's what I'm going to do.

 

[lukas]

Hi all, Czesc Marcin ty tez tutaj?
I build this preamp with good results, but i modified him: throw up capacitors C2, C5, and C10, and cancel negative feedback (R10 and C6).
This is my test results THD with oryginal channel, and mod channel:
Input 600R, 3,78mV output 600R,  0,77v(1kHz) - Normal Channel. Modified Channel:Input 600R, 2,20mV  output 600R,  0,77v(1kHz).
Oryg      THD(%)  V        Mod          THD(%)    V
20kHz    0.55      0.69    20kHz      0.47        0.71
10kHz    0.57      0.77    10kHz      0.48        0.76
1kHz      0.69      0.77    1kHz        0.50        0.77
100kHz  1.00      0.70    100Hz      0.57        0.74
30Hz      2.80      0.51    30Hz        1.20        0.68
20Hz      4.22      0.37    20Hz        2.02        0.54
Conclusion: lowest  fraquency respponce and THD on lower fraquency, More Sensitivity(more gain).
Sorry for my english

 

[bruce0]

Might get some DC  on DI and output pot, and clipping at low output pot settings?

 

[gyraf]

..you get some +100V DC offset on your input. Not recommended.