Build Thread:MS76

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I sent out a mass mailer to double check the 7912 working, and that has been verified 100%.

Got this tip back.

Check the 6.8uF tantulums. Some kits went out with 16V by mistake, and they need to be 35V, so try replacing those.

Gustav
 
I agree with Harpo, that meter looks like a plain DC meter movement. No visible signs of any internal rectifier or resistor as is normal. If it is indeed a DC meter, it will read zero until the AC signal being fed to it is clipping, which will create a DC offset voltage, which the meter will respond to. In its original application, this meter was likely driven by some active rectifier circuitry. Bottom line: just because it's scaled in VU doesn't mean it's a VU meter ... especially when its description reads that it's a 0-1 mA movement with a DCR of 50Ω as below. Surprising the 1206 tolerates such a near-short-circuit load!
Is it possible/a good idea to modify/build a circuit to get the TN-73 to work?
Otherwise is this meter Audio Level VU Meter WF1029X, 43,39 € suitable?
 
Is it possible/a good idea to modify/build a circuit to get the TN-73 to work?
Otherwise is this meter Audio Level VU Meter WF1029X, 43,39 € suitable?
According to its specs, the WF1029X is a real VU meter, namely:
"Input: The VU meter reads 0db at 1.228V +/- 10% trough an external 3.6k resistor
Internal Resistance: 3.9K ohms +/- 10%"

For decades, this is the way standard VU meters are designed (with an internal rectifier, often a miniature copper-oxide type, and a series resistor of 3.6 kΩ)
 
Is it possible/a good idea to modify/build a circuit to get the TN-73 to work?
Otherwise is this meter Audio Level VU Meter WF1029X, 43,39 € suitable?
It would certainly be possible to build an active rectifier (op-amps, diodes, etc.) to drive the TN-73, but the circuit would likely require some experimenting to achieve exact tracking with the calibrated level marks across its scale. Perhaps the schematic of a piece of gear that uses this meter could be used to duplicate the circuitry.
 
The meter is buffered, and you can probably get by by just rectifying the signal - use schottky diodes for rectification to minimize voltage drop..

Gustav
"Get by" is the operative phrase. A real VU meter has specified response times so that "VU" meters agree with each other indicating level of real program material, not just sine waves. In serious recording applications, this is important.
 
Hi everybody, I’ve been talking to Gustav via email about some issues with my MS76, and he asked if I would post in this thread in case there is helpful info for anybody else.

I’m able to complete the Q-Bias calibration on my unit, I fed it a 0.775VAC signal and tweaked the trimpot (CCW) until VU out showed a drop of 1dB. I can then zero out the VU in gain reduction mode, and try out the different ratios.

20:1 and 12:1 work as you would expect, but when I change to 8:1 or lower, the signal level drops by around 30dB, as though the unit has gone into heavy compression that it can’t recover from. The meter reflects the level drop, moving far to the left indicating heavy gain reduction. I have been measuring the DC at the 2N5457 FET, and at 2:1 ratio it measures a small positive DC voltage, even with no input signal.

I can wind back the Q-Bias control (CW) whilst monitoring the 2:1 ratio, and at a certain point the signal returns. If do this, all the ratios are compressing with normal output level, but the meter doesn’t behave correctly. It still dips off to the left when I select 2:1 or 4:1.

My unit had the 16V Tantalum caps installed, and two of them failed completely after a few hours of testing. I now have some 35V Tants in there, but that hasn’t fixed the problem.

A couple of other issues that may or may not be related. When I turn my output controls near to maximum the unit kicks out glitchy distortion, almost like ‘motorboating’ feedback. This happens with or without audio signal present.

Meter is set up, so 0VU =+4dBU.

I have to send 2.48VAC (measured at hot and cold pins of the input xlr) to get VU in to read 0VU. That's a bit hotter than +4dBu, does that suggest a problem with my meters?

Ross
 
Just finished building mine last night, but no compression happening. When I adjust Q-Bias the meter moves but doesnt respond to incoming signal at all. Some of my voltages in the CV area are way out ie. 27v not dropping down to 14v through resistor even though they are the correct value (checked).

In regards to Q1 what voltages DC & AC should we see at the DSG to verify it is working?
Cheers :)
 
I have to send 2.48VAC (measured at hot and cold pins of the input xlr) to get VU in to read 0VU. That's a bit hotter than +4dBu, does that suggest a problem with my meters?

+4dBU = 1.22V. Double, if you measure from the inverted signal.

It should be 1.22 on the 1206 output pin, iirc.

If there is an order of magnitude/reference error (will correspond to 6dB), it goes back to every units since I put it out in 2013. Never too late to correct such an error, but if so, it will not be the cause of the bigger issue.

Gustav
 
Last edited:
Just finished building mine last night, but no compression happening. When I adjust Q-Bias the meter moves but doesnt respond to incoming signal at all. Some of my voltages in the CV area are way out ie. 27v not dropping down to 14v through resistor even though they are the correct value (checked).

In regards to Q1 what voltages DC & AC should we see at the DSG to verify it is working?
Cheers :)

Hi Chad.

Could you elaborate?

The meter doesnt show incoming signal in VU in mode?

Gustav
 
I am not sure if there are a handful of builds with different problems at this point, or a handful of builds with the exact same problem.

If its the same problem - theres is, of course, as real possiblity it can be tracked down to a common component error.

So far, we have ruled out the replacment fo the 7912, as well as replacing the 6.8uF tantulums. By Linus statement, that he did actually measure all resistors when populating, it is unlikely there is a common "wrong value resistor in bag" problem. I also had Linus verify that he was 100% sure he had no wiring errors.

If its a common problem, my understanding is this.

1. You are getting signal at VU in
2. You can adjust the meter VU in
3. All voltages check out
4. All voltages printed around the transistors check out.
5. When you have the signal at 0dB at VU in, you are not able to bring up the signal to 0dB at VU out with the output pot to perform the q-bias calibration

A few ideas.

Are you 100% sure you are in LR mode, when you do the calibration? Single side signal is flakey to say the least in MS mode, so this is very important.
Have you tried checking both channels? And what happens if you jump the fet and check the signal gain throughout the input and line-amp "solo" so to speak?

Gustav
 
Last edited:
+4dBU = 1.22V. Double, if you measure from the inverted signal.

It should be 1.22 on the 1206 output pin, iirc.

If there is an order of magnitude/reference error (will correspond to 6dB), it goes back to every units since I put it out in 2013. Never too late to correct such an error, but if so, it will not be the cause of the bigger issue.

Gustav

Hi Gustav, yes 2.48VAC between input pins 2 & 3 gives me 1.228 out 1206 output pin (referenced to ground) and 0VU. Looks like the meter is 6dB down. As you say though, not too important at the moment
 
I'm having different issues to everyone else Gustav.

1. You are getting signal at VU in
2. You can adjust the meter VU in
3. All voltages check out
4. All voltages printed around the transistors check out.
5. When you have the signal at 0dB at VU in, you are not able to bring up the signal to 0dB at VU out with the output pot to perform the q-bias calibration

I can bring the signal up to 0dB

Are you 100% sure you are in LR mode, when you do the calibration? Single side signal is flakey to say the least in MS mode, so this is very important.
Have you tried checking both channels? And what happens if you jump the fet and check the signal gain throughout the input and line-amp "solo" so to speak?

Definitely in LR mode. If I remove the FET unit operates as a line amp with input and output gain

I can perform the Q-Bias calibration - it's very touchy though!! The 1dB drop is achieved with the slightest fraction of a turn of the trimmer.

When I switch to the lowest ratios, the unit moves into heavy compression with or without an input signal. The meter moves far to the left, the signal level (if I am feeding it) drop around 30dB, and the DC voltage at the gate of the FET moves to a positive DC voltage.

So far, we have ruled out the replacment fo the 7912, as well as replacing the 6.8uF tantulums. By Linus statement, that he did actually measure all resistors when populating, it is unlikely there is a common "wrong value resistor in bag" problem. I also had Linus verify that he was 100% sure he had no wiring errors.

I also measured all components before stuffing, and have since verified all of them using the PCB layout you sent me. Particularly the ratio board resistors.

Could the FD333 diodes have been damaged if the 16V Tantalum caps failed?
 
I can perform the Q-Bias calibration - it's very touchy though!! The 1dB drop is achieved with the slightest fraction of a turn of the trimmer.

Perfectly normal, that its touchy, - also why I added the multiturns at a revision of the board. It sounds like its almost as touchy as it used to be with the non-multi turn on yours, though. Ive calibrated a lot of these, so even if its an odd/non reliable reference, just closing my eyes and feeling it out, Id say you should see about 30 degree movement on the 1dB travel.

Could the FD333 diodes have been damaged if the 16V Tantalum caps failed?

Had to check, and it looks like their reverse breakdown is rated 150V, so I doubt it, but its worth a try for a relatively easy swap/low cost part.

Gustav
 
Fixed my problems today so thought Id come back and update.
1. My voltages were out because I used a 270k instead of 270R, watching tv while stuffinrg resistors is a double edged sword. Checked them in circuit around the area where voltages were too far from the silkscreen values.
2.I assumed (wrongly) that the Gain Reduction mode would be engaged without the GR relay being energised (normalled ON). Make sure you have your switches installed and relay power running when you test.
3. A few solder pads came off during heating, not quite as sturdy as some PCBs for removing components so go slow.

Hope that helps someone, mines all working sweet now. Sounding great :)
 
Hope that helps someone, mines all working sweet now. Sounding great :)

Glad to hear it - one down!

Stick around, if you dont mind - can check a few voltages with you to get the last guys over the finish line. (I dont have a unit around).

Gustav
 
Stick around, if you dont mind - can check a few voltages with you to get the last guys over the finish line. (I dont have a unit around).
Haha except us poor fools who were too late to the party and ended up buying one of the leftover boards but no kit! Does a BOM exist for this, or was that not published since it was mainly sold as a full kit? I won’t be able to start on mine for awhile yet, but I’m just thinking ahead a little.
 
Haha except us poor fools who were too late to the party and ended up buying one of the leftover boards but no kit!
I was refering to the people who had posted problems, that we were investigating right now. If you have problems, when you finish your build, the thread will still be here, and you can still ask for help (unless the forum closes).

Sorry, I dont have a BOM, for the reason you stated - youll have to go over the board or schematic.

Gustav
 
Thanks Gustav! I figured as much regarding the BOM.

I am hoping to get to this build this autumn (end of summer and early fall are incredibly hectic for me). Can’t wait! I have to say, it’s really admirable that you give such great support here.
 
Back
Top