All things G1176 - the new "repost" thread.

[dustbro]

[quote author="gswan"]No, I use the unit's meter to do Qbias calibration. However my switch arrangement is different, since I have the GR disable switch on my attack pot.[/quote]
That'll do it

 

[Siegfried Meier]

So, can anyone figure out the strange noise/buzz that's happening in both my units? It can't be toroid related, as it changes with the positioning of the input and output knobs.

I've posted an example here of what it sounds like, and this is with me turning the input and output knobs. I've boosted the gain to full so you can hear it (watch your levels). Can anyone guess what's wrong?

http://www.siegfriedmeier.com/electronics/1176 Noise.mp3

Thanks,
Sig

 

[gswan]

Sounds like power supply buzz to me. How clean are your power rails?

 

[Siegfried Meier]

How do I check how clean the power rails are?

Thanks,
Sig

 

[gswan]

If you have a CRO you can use it to measure the noise and ripple on the supply rails. Without a CRO you're kind of stuck.

Instead you might also like to check your electro's around your PSU. What PSU are you using? Check that they are all in correctly and that they are the correct value. How warm does your regulator get?

 

[Siegfried Meier]

What the hell is a CRO? I'm sure it's a meter of some type that I don't own.

I used the P/S that was built in to the Mnats Rev J boards. I'll check the electro's, but the regulator doesn't get too hot at all - I can definitely touch it without getting burnt. Keep in mind that both units do the exact same thing, so it's something I've done to both.

Thanks,
Sig

 

[gswan]

LOL. It's a piece of test equipment called a 'cathode ray oscilloscope'. Probably called a scope these days, as the new ones are LCD and LCDO isn't easy to say. It's kind of essential to track down tricky faults on these kinds of things, since we need to identify the source of the noise.

OK, so your regulator doesn't get hot - that's a good thing. Check the PSU for the +30V rail, since if it has a lot of ripple or noise on it it can find its way to the output.

It could also be pickup from cables near the mains wiring - I don't know what your assembly looks like so it's just a guess. What are you running through the power switch on the pushbutton assembly?

Did you do the test with the audio file with the input terminated at all? Best thing to do is to short the input to ground when doing this, as we don't want stray signals influencing the test results.

 

[Siegfried Meier]

Ahhh of course a scope, that makes more sense to me haha.

I'll check the PSU +30 rail again. I'm taking the hot from the input and sending that to the power points on Mnats pushbutton boards, and then going to the Toroid in. Here's what it looks like on the inside:

Input XLR has the shield going to the chassis. I don't really have any wires near the mains switch, besides the ones from the meter pushbutton board, and that's pretty hard to eliminate.

Thanks,
Sig

 

[gswan]

All the wires running to the ratio and meter switch board can pick up noise. Particularly from cables with large AC voltages on them. I never liked the idea of have 120 or 240V AC a few millimetres away from signal wiring from the output on the meter switch.

Set your pots so you get maximum noise out (with shorted inputs and balanced outputs) and then try moving the secondary cables from your toroid. I notice that they are quite long and run parallel to some other cables in there.

See if the noise varies in amplitude when you move these closer or futher apart. If so then you may have pickup from the transformer secondary, the solution being to move the toroid over and shorted these cables, or attach the toroid to the rear panel and run the cables along the rear panel.

 

[Siegfried Meier]

Alright, I tried moving around all the cables, I even tried moving around the Toroid, turning it, moving it to the back, nothing at all. The noise remains constant and doesn't change.

I also tried a cheater plug to remove the AC ground, and the low, low hum went away, but that annoying higher buzz still remained. This is really starting to bum me out. I rechecked all the voltages, and they remain steady and don't waver at all with a simple DMM. What else can I try?

EDIT - it's really strange, but the noisiest level is when the input pot is all the way down, and the output pot is all the way up. You'd think the input would be up all the way too.

Sig

 

[gswan]

So you've got a few grounding issues to sort out to eliminate the buzz.

The higher sound may also be reduced by this, however moving cables around didn't affect it I'd guess that it's not induced by these cables.

Your pot positions would indicate noise pickup around the amplifier stage formed by Q2/Q3/Q4 or it's being being picked up and transferred to the top of the pot from the ratio switch or the GR disable switch.

First thing I'd try is to remove the AC mains power wiring from the switch module and tape them together away from the front panel, just to see if it makes a difference (since this could transfer noise to the top of the output pot).

 

[Siegfried Meier]

Ok, cut the AC lines from the pushbutton switch, taped them together and kept them to the back of the chassis - no change in noise.

I also checked those red and black wires that go from 21 and 22 of the main board to 21 and 22 of the Ratio board. I disconnected them, and the noise got a million times better. What are these wires? It says they don't need to be shielded wire according to Mnats wiring guide, so I've just left them as stranded wires, and twisted them as he has shown. Could these be a problem?

Sig

 

[gswan]

OK, that's good. So we're not getting noise from the mains AC input on the power switch board.

I'd certainly shield cable 22, since it is the input to the GR amp and indirectly connects to the top of the output pot. Cable 21 is a bias voltage which is summed with the GR rectifier output.

Check your shielding scheme as well, don't just blindly connect both ends of the shield (this can cause gound loops).

 

[dustbro]

crap... I'm driving myself nuts with this thing. When I turn the Qbias, I hear the signal attenuate and I see the meter deflect. but it stays there... no recovery when I pull the input signal. It's like the darn thing has the ability to compress and had the ability to meter the GR, but it's not affected by the input. After trying the calibration several times I was able to make the meter bump up (the wrong way) about 1db when I crank the input.
But I guess the thing is compressing. When I go to calibrate the Tracking, I'm able to measure the input and output voltage and it does get to the point where the output is 1/2 the voltage of the input... which means it must be attenuating the volume. But it's barely audible. Nowhere near as much as if I crank down the Qbias pot. I think I recall the input voltage being around 5V and the output around 2.5V. Does that seem out of whack?

 

[gswan]

When I turn the Qbias, I hear the signal attenuate and I see the meter deflect. but it stays there... no recovery when I pull the input signal. It's like the darn thing has the ability to compress and had the ability to meter the GR, but it's not affected by the input.

It's not supposed to recover when you remove the input signal, it's a 'bias' setting. You have the GR disabled, so there is no signal affecting the FET conduction, just a tiny bit of bias to start conduction and put it into the linear conduction region.

What you are seeing by measuring the 1dB drop is the amount of bias being applied to the FET reflected as a signal attenuation.

There's no point calibrating the meter tracking until you have the GR FET properly calibrated first, since you will be tracking against the wrong thing.

 

[Siegfried Meier]

None of the shields are connected to both ends. I've very strictly followed Mnats wiring diagram instructions. 22 is shielded already, as shown in the wiring guide. I was inquiring about 20 and 21, which are just stranded and not shielded. I assume those are ok. What else to try?

 

[Siegfried Meier]

Ok, did some more testing today. I was able to reduce some more hum by either connecting the Input XLR shield to the shield of the input pot, or to the shield of the output XLR - same amount of hum reduced in both instances. But, it's still not good enough - that weird P/S buzz still remains and drives me crazy. I'll keep picking away at it.

Sig

 

[dustbro]

[quote author="gswan"]
What you are seeing by measuring the 1dB drop is the amount of bias being applied to the FET reflected as a signal attenuation.

There's no point calibrating the meter tracking until you have the GR FET properly calibrated first, since you will be tracking against the wrong thing.[/quote]
Right right... I think I jumped the gun on my timeline. Here's what I did.
I am able to complete the Qbias calibration. I used an external meter since the G1176 shows no meter movement when bypassed. I am able to adjust the Qbias pot CW to make the signal (1K, 0dB from tone generator) drop by 1dB. Then I move over to the Tracking adjustment. I put a mulitmeter across pin 1 and 3 (unbalanced from generator) and read the AC voltage. I put a 2nd meter across pin 2 and 3 of the output and read the voltage. I adjust the output from the generator (not input knob on 1176) until the output voltage of my 1176 is half of what I'm putting into it. In this case the input voltage was 5V and the output was 2.5. Then I adjust the Tracking pot until the meter reads -6VU. The problem is that even when I crank the tracking pot full, I can only get the meter to drop to -1.
If I tweak the Qbias pot a tad I am then able to adjust the tracking to get the -6 needed for the adjustment. I can then set my GR meter to zero, but no matter what level I input, the meter will not show any GR. If I really crank up the adjustment pots I can get the meter to bump in the wrong direction when I input a super loud signal.
I appreciate all of the suggestions so far!

 

[Siegfried Meier]

Ok, so I've finally settled on connecting a new ground cable from the shield of the input pot to the shield of the output pot - this reduced the hum by the most amount. If you crank things up, there's still a small amount of that buzz, but within reasonable working levels, this did the most. Did anyone else have to connect these 2 shields together with a lundahl in and out Mnats Rev J 1176?

Any other grounding tips would be appreciated.
Thanks,
Sig

EDIT - also connecting another shielded cable from the output XLR to the input pot (or output pot) shield further reduced noise by a small amount. It's funny, because these are all things that everyone says NOT to do.

 

[gswan]

[quote author="dustbro"]
Right right... I think I jumped the gun on my timeline. Here's what I did.
I am able to complete the Qbias calibration. I used an external meter since the G1176 shows no meter movement when bypassed. [/quote]

Try pressing the +4 and GR disable switches together. You should then be able to measure it on the unit's meter.

[quote author="dustbro"]
I am able to adjust the Qbias pot CW to make the signal (1K, 0dB from tone generator) drop by 1dB. Then I move over to the Tracking adjustment. I put a mulitmeter across pin 1 and 3 (unbalanced from generator) and read the AC voltage. [/quote]

With a transformer input you can put your generator straight across the transformer, ie pins 2 and 3.

[quote author="dustbro"]
I put a 2nd meter across pin 2 and 3 of the output and read the voltage. I adjust the output from the generator (not input knob on 1176) until the output voltage of my 1176 is half of what I'm putting into it. In this case the input voltage was 5V and the output was 2.5. Then I adjust the Tracking pot until the meter reads -6VU. The problem is that even when I crank the tracking pot full, I can only get the meter to drop to -1. [/quote]

Not the procedure I use. Something doesn't sound right here. Try this instead.

- No signal input, GR disabled, metering GR (ie press both switches), ratio 20:1. Adjust R55 (zero adj) until the meter reads 0VU.
- Input halfway, output fully CW, release fully CW, GR disabled, metering +4 (press both switches again). Apply -10dB to the input (0.245Vrms).
- Turn the ouptu control until the meter reads 0VU.
- Enable GR (release the GR disable switch) and adjust the input control until -10 is indicated on the VU meter.
- Disable GR and re-adjust the output level control for 0VU.
- Repeat these steps until you get -10 on the meter when GR is enabled.
- Leave the input and output controls now and switch the meter to GR.
- Adjust R54 (tracking) until the GR meter reads -10 when GR is enabled (ie releasing the GR disable switch).
- Disable GR and adjust R55 until the meter reads 0VU.
- Repeat these steps if necessary so that the GR metering shows -10 when GR is enabled.

[quote author="dustbro"]
If I tweak the Qbias pot a tad I am then able to adjust the tracking to get the -6 needed for the adjustment. [/quote]

You don't want to touch the Qbias pot once it is calibrated.