The official G9 help thread

[glover]

Any tips ?

 

[bruce0]

Any tips ?

Are you using a balanced connection to the mixing desk?

 

[MeToo2]

Any tips ?

To be honest, the structure looks perfectly normal (some 100Hz rectification noise and some high frequency hiss). Without seeing a calibration scale and gain structure of how loud/quiet it is relative to your signal it's hard to know if it's a problem or not. Can you calibrate the sound card e.g. with a loopback and a multimeter, and then measure a sine tone from your DAW relative to the noise?

 

[Spencerleehorton]

Hi Guys,

Have test my Gyraf G9 and all but one thing seem good, on 8-11 gain settings i get massive oscillation on both channels.
I dont have cans on OEP transformers or screened wires yet as wanted to ask for up to date info for doing all this as have looked through pages and pages of threads and havent found a definitive answer yet.
If you could help with this please, also im using Dragon E82CC valves and wondered whether there are any considerations for these or other valves that can be used that are good?

regards

Spence.

 

[glover]

Yes, I'm using balanced connections. MeToo2, I'll try to calibrate it and send screan.

Spencerleehorton - as you might expect you need to mount screening cans on OEP's and apply DI mod - connect front jack's direct to main pcb here http://www.gyraf.dk/gy_pd/g9/G9-EDIT.GIF you'll find info about this mod. And my advice is to use as much screened wires as you can. ;D Of course in signal lines 

 

[Spencerleehorton]

HI Glover,

thanks for that, i will start implementing all of this and let you know the results.

regards

Spence.

 

[glover]

Ok MeToo2, what I have done:
1 - i connect my interface output 3 with input 3, send 1kHz RMS test tone at -18dBFS and trim interface gain on input 3 to have -18dBFS in pro tools so I get unity gain on +4dBu.
2 - disconnect output 3, to input 3 i connect output from G9, at G9 input i connect dynamic microphone with switch to get load on input (of course mic was disabled). Then I record track with input and output on 0 (on G9) and on max - didn't move interface gain.
3 - on left channel which is further from PSU block I get:
all knobs on 0 ->        -75dbFS of noise
all knobs on max ->  -63dBFS of noise
And on right channel (closer to PSU) I get:
all knobs on 0 ->        -75dbFS of noise
all knobs on max ->  -56dBFS of noise
Both channels set to MIC.
Is this acceptable score ? Still waiting for new tubes to check this possibility.

 

[museic]

Could someone help me out with a mouser part number for the 10k trimmer that sets the 48v?

The one I got doesn't fit in between the caps.

 

[glover]

Here you go -> Mouser Part No: 531-PT10LV10K, look like this -> http://www.tme.eu/pl/details/pt10lv-10k/potencjometry-tht-jednoobrotowe/piher/pt10lv-10k-a2020/#

I use this one and it fit's but I don't know what caps you have.

 

[MeToo2]

Ok MeToo2, what I have done:
1 - i connect my interface output 3 with input 3, send 1kHz RMS test tone at -18dBFS and trim interface gain on input 3 to have -18dBFS in pro tools so I get unity gain on +4dBu.
2 - disconnect output 3, to input 3 i connect output from G9, at G9 input i connect dynamic microphone with switch to get load on input (of course mic was disabled). Then I record track with input and output on 0 (on G9) and on max - didn't move interface gain.
3 - on left channel which is further from PSU block I get:
all knobs on 0 ->        -75dbFS of noise
all knobs on max ->  -63dBFS of noise
And on right channel (closer to PSU) I get:
all knobs on 0 ->        -75dbFS of noise
all knobs on max ->  -56dBFS of noise
Both channels set to MIC.
Is this acceptable score ? Still waiting for new tubes to check this possibility.

It's still not clear to me exactly what your -18dBFS is calibrated to in the real world. [dBFS = relative to Full Scale = interface dependent and pretty meaningless in the real world]

Are you saying that -18dBFS = +4dBu = 1.227V RMS into 600ohm load?

And your noise floor is at -75dBFS -- -18dBfS +4dB = -61dBu with all volume controls turned down? and -42dBu with all gain controls @ max?

If that's true that's poor. So I'd really question your calibration. Or how clean your input termination is.

 

[bruce0]

Quite possible that the mic "switch" just opens the circuit and does not leave a load on.
(I have a little XLR male shell for that purpose. It is just a XLR-M with a 150R to 200R resistor strapped between pin 2 and pin 3.  You don't even need the shell, just 2 XLR pins and a resistor would be fine if you keep the lead short (it is low impedance it is not going to pick up any noise))

I end up with lower noise figures than you are mentioning.  (Update:  The discussion that followed on noise floor testing makes me question my numbers... The noise floor of my current (lousy) audio test setup is about -95db (measured with RTAA and I don't know how it is weighted) and i see noise in the -87 to -89db range with the G9 in the loop at gain position 4 (that is from my memory ... memory with just some old tubes If you want me to make a real measurement I will.)).

So on to your problem...
The fact that you can hear it on the input at low gain settings indicates a serious problem. 

Can you tell us exactly what desk (part number) it is getting plugged into?

I looked at some of the phonic mixers online. It is not 100% clear that they are accepting balanced inputs (the use the word balanced, but they word it very carefully when talking about the line inputs (accepts input from TRS or TS).  The send return input is obviously unbalanced.

If this is a ground loop problem you can test it as follows:
a) Connect some good ground G9 to some good ground on your phonic mixer.  This can be a wire and a pair of alligator clips if no ground post is available.
b) make sure all three of those components (test equipment, signal generator etc, G9 preamp, phonic mixer) are plugged into the same plug strip.

With that does make your test again, and if that changes anything you have  grounding problem.  (The opposite is not true,  if it doesn't change anything you may still have a grounding problem).

Things to check if the Phonic desk really does have a balanced input
Suggest:
1) Assure proper termination of the mic input
2) Assure proper grounding of the XLR connectors and the box
3) Assure proper BALANCED connection from output to board
4) Assure a proper function BALANCED cable with XLR / TRS connections and not TS connections (us a DVM for continuity tested on pin 2 and pin 3.  Also test to make sure that the pin 1 is not connected to the shell or to either pin 1 or pin 2 (people make the weirdest cables).
5) Assure that the input on your board is actually balanced or just "accepts" balanced connections.

 

[glover]

Someone please tell me one more time how to accurately calibrate interface to measure noise floor.

 

[Spencerleehorton]

Hi All,

Have got my G9 all fired up and working, little bit noisy (hiss), gonna go through it all and see about testing noise floor.
Just wanted to know with the gain full up should it go into quite big distortion or not?
Mine is clean as a whistle which i was thinking it would sound a lot more pushed.

regards

Spence.

 

[gemini86]

The G9 is very very clean. It does not clip under normal circumstances. There are ways of making it dirty, if that's needed.

 

[gyraf]

Yes, the G9 has a very gradual transition into overload.

It's not that it's overly clean - it just doesn't break up suddenly.

This is what we like at Gyraf - all our units are designed to behave like this.

Jakob E.

 

[Spencerleehorton]

Hi GUys,

Just summed a mix through my G9 and apart from the noise floor being noticeable, its a lovely punch fat sound.
Any tips on testing the noise floor or has this already been covered?

regards

Spence.

 

[gyraf]

On max. gain and trim levels, no pad, front panel jack shorted, you should have some -60dBm of noise - ca. 1mV

With the ca. 60dB gain of the unit, it gives a noise figure around -120dB

Jakob E.

 

[Spencerleehorton]

Hi Jakob,

will try and test mine and see what numbers come out!!
sounds really good though, definitely now within a bit, noise wise. ive summing a mix, put bass guitar through there and sounds crisp and fat (phat!!!) i dont have the cans on the transformers yet as wanted to test it without and then put the cans on to see if there was any improvement, the biggest improvement was screened cable on the jack input, output pot etc.
Would love to make a 8-16 channel desk with all this in, but hey i can dream!!!!

Im wanting to make a 16/32 channel summing mixer for all these pre's ive now build with the facility to direct out of each channel to DAW, do you have anything or know anything suitable?

regards

Spence.

 

[MeToo2]

Someone please tell me one more time how to accurately calibrate interface to measure noise floor.

0dBFS is  the maximum an interface can output or input before clipping. It's handy for software and digital, but not much use for analogue measurements. dBFS can be different for input and output, even on a single interface.

To calibrate to real world voltages: Set up a tone generator in your Pro tools. Set it to 50Hz. Set it to some nominal level. Set the interface to +4dBm operation if you can. Put a small load on the output of the interface (say 600 ohms for the sake of argument, but most modern gear is much lower output impedance than that). Measure the voltage across the load using a True RMS AC multimeter or oscilloscope.

0dBm (600 Ohms) = 1mW into 600 ohms = 0.7746V AC RMS. Studio level of +4dBu is 1.228V AC RMS.

Either adjust your nominal level of your tone generator to get exactly 0.7746V, or alternatively use the following calculator to work out what level of dBFS(out) matches to dBu. So say you had -10dBFS on your tone generator output giving you 1.4V AC on your multimeter, that means -10dBFS corresponds to 5.14dBu. or 0dBFS = 15.14dBu.

http://www.sengpielaudio.com/calculator-db-volt.htm

Then you have to repeat the exercise using a loop back cable to the input (suggest setting your input trim to Max) to see what that calibrated level of output measures on your input in terms of dBFS(in) using a level meter in Protools.

 

[Spencerleehorton]

Hi Guys,

im not very technical with what has been said but will pass this along to a friend who is helping me out and removing all the fog on these issues:

this is what he has said about testing the noise floor and would be interested in what you think:

for meaningful results, some complex filtering and weighting has to be applied.

It is not clear(or is it?) that Jacob used any filtering or weighting. If he did, then we can't reproduce his test. If he did a straight measurement of wideband un-weighted noise then I can provide the equipment to do the same.

But beware of taking too much notice of a result using this method and certainly don't quote the figure obtained that way in any technical item as you will be shot down in flames.

does this make sense?

regards

Spence.