Calibrating G1176

[byoung]

I'm trying to calibrate the unit right now....this may seem like a dumb question.... ok very dumb...but I've got to ask.....when attching the DMM to the inputs or outputs on the balanced lines.... I attach the leads to + and 0, or - and 0 or across + and - ? I knew this at one time but can't find what I did with my notes. I've been doing + and 0 and when adjusting the GR meter tracking 50mV In is nice and stable but I can't get a stable 50mV on the outputs, it's jumping around 48-59mV without me doing anything, which has led me to think I've got the DMM hooked up wrong.

EDIT: OK I went through the process and followed Gyrafs instructions using my DMM on + and 0 leads. Now the last step was to reset GR meter zero so I did that. And now running the same signal that I just adjust as -6db, the meter is only at -2 on the VU...is this correct?

 

[stefan]

I´m confused about the "QBias" adjustment, what does "adjusting the trimmer until a level drop of 1dB is achived" mean? Should it say "start with the trimmer fully CCW" or "1dB drop when the unit is turned on" or something like that? :?:

 

[ChuckD]

For the 50mV on the GR meter Trck I assume this is DC?

Does anyone know how to get 50mV out of a DAW? Using protools and a 002 rack A/D interface but I don't see 50mV on the outputs. I can only measure VAC and not DC.


Thanks

-ChuckD

 

[gswan]

[quote author="stefan"]I´m confused about the "QBias" adjustment, what does "adjusting the trimmer until a level drop of 1dB is achived" mean? Should it say "start with the trimmer fully CCW" or "1dB drop when the unit is turned on" or something like that? :?:[/quote]

Qbias adjusts the DC negative voltage that is applied to the FET gate (both the GR and metering FET) to turn them slightly into conduction (ie they are just starting to conduct).

Essentially you need to start with them turned off (setting the Qbias trimmer fully CCW), apply a 0dB 1kHz signal and adjust the input until you read +1dB on your meter (set to +4dB metering). You output pot needs to be fully CW to get the meter to read.

Now adjust the Qbias pot so that the FET is just starting to conduct - this will appear as a drop in the signal.

You need to adjust it for a drop of only 1dB. If you use a single turn pot, the turn on knee can be quite fast so only a slight tweak on the Qbias pot will result on a large signal drop.

There are advantages to matching the FETs, and I put some notes up on my page here about this (http://www.axtsystems.com/1176clone/)

 

[gswan]

[quote author="ChuckD"]For the 50mV on the GR meter Trck I assume this is DC?

Does anyone know how to get 50mV out of a DAW? Using protools and a 002 rack A/D interface but I don't see 50mV on the outputs. I can only measure VAC and not DC.


Thanks

-ChuckD[/quote]

What would you want 50mV for?

Once you've done the compression setup using the -10dB input signal and managed to get a 0dB->-10dB drop when switching on the attack pot you need to adjust the GR trimmers to also show this on the meter when it is switched to GR.

You want to see a -10dB reading when the bypass is off and 0dB when the bypass is on (ie no compression). This is done using the R54/R55 adjustments.

 

[ChuckD]

I have a question on how the QBias or DC ref voltage for the FETs works.
I recently re calibrated the unit and I am finding the ratio knob no longer changes the amount of compression... well not enough to notice really.

Does this mean I am to far into conduction, or not in conduction enough? In other words have I calibrated the fet with greater then -1dB or less then -1dB... Would this cause the effect I am seeing?

I am guess it is too far into conduction or greater then -1dB reduction which would move me beyond the knee and all compression would be simular?

Anyone want to help me better understand this?

-ChuckD

 

[gswan]

[quote author="ChuckD"]I have a question on how the QBias or DC ref voltage for the FETs works.
I recently re calibrated the unit and I am finding the ratio knob no longer changes the amount of compression... well not enough to notice really.

Does this mean I am to far into conduction, or not in conduction enough? In other words have I calibrated the fet with greater then -1dB or less then -1dB... Would this cause the effect I am seeing?

I am guess it is too far into conduction or greater then -1dB reduction which would move me beyond the knee and all compression would be simular?

Anyone want to help me better understand this?

-ChuckD[/quote]

First check to see that the GR fet (Q1) is actually in conduction by going through the Qbias calibration stage.

How are you measuring that the 'ratio knob no longer changes the amount of compression'?

The amount of compression is set by two divider chains on the ratio switch.

One divider chain alters the amount of signal (taken prior to the output control) that is fed into the GR amp. The GR amp produces both inverted and non-inverted outputs which are AC coupled to the rectifiers. This means that GR is detected on both + and - excursions of your input signal.

The second divider chain on the ratio switch provides a negative offset which is summed with the AC coupled signal prior to the rectifiers. This has the effect of changing the signal threshold required for compression to take effect.

Perhaps your input signal level is below the threshold? Try measuring the compression ratios using a signal generator and dB meter. Don't rely on the meter on the unit until you are sure that it is calibrated and tracking correctly.

 

[ChuckD]

Mainly I was relying on my ears to test the amount of compression for each ratio setting.

Before re-calibration there was a very noticeable difference between them then I buggered it up by re-calibrating as my GR meter was never working properly. I should have just left it alone, as I never really use the meter...

Try measuring the compression ratios using a signal generator and dB meter. Don't rely on the meter on the unit until you are sure that it is calibrated and tracking correctly.

Where do I take these measurements?

Also do you know if What I asked before is true that I am too far in conduction or not far enough?

Thanks!
-ChuckD

 

[gswan]

Chuck,

Relying on your ears is probably not a good idea for calibration purposes :grin:

If you buggered it up by re-calibrating, and you GR meter was never working properly then it probably means that there are some construction errors or faults or problems with your test and measurement procedures.

If you have the FET too far into conduction then it will be attenuating your signal all the time, even with GR bypassed. The attenuation will be constant in bypass mode however so you can make up the gain loss in the output stage.

However when you turn on the GR amp and start applying signals to the FET, it is possible that the FET could move out of the linear region and into the saturation region, resulting in no further discernable gain reduction regardless of the input signal level.

If you look at the bottom FET measurement chart here (http://www.axtsystems.com/1176clone/fets.html) you can see that the FET has a reasonably linear characteristic between conduction and saturation. The initial Qbias setup puts it just into the conduction region and signal excursions operate over the linear portion of the curve.

To measure your compression ratios, feed a 0dBu signal into your compressor with the input control about halfway, release CW, attack CCW and adjust the output control (with GR bypassed) to give you a 0dB reading at the output. Then switch on compression. Don't touch the input or output gain knobs and measure the output signal level for a range of inputs (say 0db to -15dB) for each ratio.

I'm about to put a page up on my site to describe the test, however the results for my dual unit are:

Ratio  Channel A  Channel B
  4      4.2         5.5
  6      5.4         6.7
  8      7.0         8.6
 10      8.0        10.0 
 12      9.2        12.0
 20     13.6        20.0

You can see quite a difference between the marked settings and the actual ratios and a difference between the channels. I shall probably try to tweak the settings for channel A to closer match the panel markings at some stage.

 

[TornadoTed]

My Q Bias rimmer is way too sensitive. It does nothing and then suddenly it's compressing massively, it even changes sometimes just from removing the screwdriver!

What can I replace the trim pots with that will give me a larger 'sweet spot'?

 

[gemini86]

[quote author="TornadoTed"]My Q Bias rimmer is way too sensitive. It does nothing and then suddenly it's compressing massively, it even changes sometimes just from removing the screwdriver!

What can I replace the trim pots with that will give me a larger 'sweet spot'?[/quote]

a multi-turn trimmer...bourns work really well.

 

[TornadoTed]

Sorry to be a bit stupid.

Something like this work?

http://uk.farnell.com/9353577/passives/product.us0?sku=bourns-3296-1-202lf

 

[gemini86]

[quote author="TornadoTed"]Sorry to be a bit stupid.

Something like this work?

http://uk.farnell.com/9353577/passives/product.us0?sku=bourns-3296-1-202lf[/quote]

that's the one...

 

[TornadoTed]

Thanks for your help Gemini86.

I have been using these for ages without properly calibrating them. I don't mind using my ears and forgetting the meters exist but I have outside producers coming in soon and I would like to get them right.

 

[JW]

Hey guys,

Is there any reason I can't just use external VU meters for both the Q bias and meter tracking calibrations?

For Q-bias, I just used the vu on my console (while the 1176 is in bypass mode) to see when the q-bias is adjusted (or over-biased as they might say in the tape machine days) so it drops 1dB. Is there anything wrong with this?

And concerning the meter tracking, it's a lot easier to just see on my VU meters when the output is half the voltage of the input than trying to track down a second DMM that may or may not have true RMS on it to take exact voltage measurements of both input and output. I know it's not as exact, but does it need to be?

 

[gswan]

Hey guys,

Is there any reason I can't just use external VU meters for both the Q bias and meter tracking calibrations?

None whatsoever for output signal level measurements. The VU meters of the 1176 are connected across the outputs, so you are measuring the same thing. GR metering relies on internal signals, so can only be displayed on the internal meter.

It may not be as accurate as using a scope or decent measuring device since the VU meter is unlikely to be particularly accurate nor have a good measuring range or resolution. Qbias should be setup as accurately as possible, since it has a significant effect on the way the compressor operates.

If you attach a meter to the input and a meter to the output and plot the transfer characteristic for various ratio settings you will be able to see where the knee is, how soft it is and measure the actual compression ratio value.

 

[JW]

Okay,

Well, concerning the Q-Bias then, I used this equation posted by Dean on the first page of this thread:

db = 20[log(V1/V2)]

To get the RMS voltage equivalent of 1DB, and using .775 for V2, and solving for V1 I'm getting .87V's
So, with a .775V signal inputting, I should adjust the Q-Bias pot for (.775V-.87V) = -.095V

Correct?

 

[gswan]

Okay,

Well, concerning the Q-Bias then, I used this equation posted by Dean on the first page of this thread:

db = 20[log(V1/V2)]

To get the RMS voltage equivalent of 1DB, and using .775 for V2, and solving for V1 I'm getting .87V's
So, with a .775V signal inputting, I should adjust the Q-Bias pot for (.775V-.87V) = -.095V

Correct?

Doesn't sound correct to me. You don't appear to understand what you are doing.
For a 1dB drop from 0.775Vrms the voltage at the output will be 0.691Vrms, compared with the 0.775Vrms at the input.
Your VU meters will be reading in dB so you can read a 1dB drop directly off them, rather than calculate the voltage.

 

[JW]

Hi gswan. Thanks. You're quite correct when you say that I don't know what I'm doing.

Anyway, whether or not it's okay to do this calibration by using my mixer's VU meter after the output of the 1176 was pretty much what I was asking.

If it needs to be more exact than that I can measure the output voltage, but I'm more confident in just tweaking the Q-Bias for a dB drop while looking at my console VU.

 

[gswan]

That's what the calibration step is for, to place the GR FET just into conduction. Calibrating it with external accurate VU meters is OK, however if you have better test equipment at hand then use it instead.

This step starts off with Qbias wound out (so the FET is completely off). Run a signal through the amp stages so that you are getting a 0dBu reading at the output and then carefully wind up the GR FET bias until it is just conducting and you get a 1dB drop in the output.

After this you can check your compression ratios to confirm that you have the correct components in your GR amp.