Peak Indicator Light causing noise/click on console?? Help!

[galensjp01]

Hey everyone!

Need some help diagnosing a strange problem. I recently picked up an old Philips Mix 5 small format mixer. (Schematic attached)

It sounds great and is running really quiet but I’m running into one weird issue. It has a double peak indicator light for each channel (and ones for the output channels and aux). The 1st light is green and then it shifts to a second light (red) when that channel is clipping.

The issue.. there’s an audible “click” or “pop” sound when the light changes from red to green. It happens as part of the decay after the initial transient. It’s weird and the timing of them can be different so sometimes there’s a water falling of click sounds. It’s hard to explain.. maybe I’ll take a synced video and show you. But yea.. any idea what could be causing that? Is voltage from the peak light system leaking into the audio path somehow??

Thanks!

 

[Script]

Peak detector is on p. 38 of the pdf.

Looks maybe too simple as a circuit?

Could try and disconnect D301 to take it out of circuit. Then send same signal into unit as before to check whether that 'noise' really come from there.

Could be, the red LED is tuned to an absolute ceiling, meaning if it lights up, the circuit is clipping elsewhere. Most peak indicators I've come across are buffered (op amp), and they light up with a headroom margin. Maybe this one doesn't?

Also not sure whether hooking the diode rectifier of the limiter (to opto in feedback path) unbuffered onto the audio path is a good idea ?

 

[galensjp01]

Peak detector is on p. 38 of the pdf.

Looks maybe too simple as a circuit?

Could try and disconnect D301 to take it out of circuit. Then send same signal into unit as before to check whether that 'noise' really come from there.

Could be, the red LED is tuned to an absolute ceiling, meaning if it lights up, the circuit is clipping elsewhere. Most peak indicators I've come across are buffered (op amp), and they light up with a headroom margin. Maybe this one doesn't?

Also not sure whether hooking the diode rectifier of the limiter (to opto in feedback path) unbuffered onto the audio path is a good idea ?

Appreciate the reply! I’ll try to remove it from the circuit and see if the noise disappears and report back!

Is there a do-able way to add an opamp to the peak circuit?

 

[JohnRoberts]

Look at the layout, the LED responding to peaks causes a current spike from somewhere to somewhere. Look at where the current is flowing. Perhaps slowing down the rate of change with a feedback cap, perhaps use a cap to manage the current spike.

Sorry I didn't look at the schematic yet.

[edit- FWIW I always flashed the peak LED 3dB before clipping with an added hold circuit to make narrow peaks easier to see. /edit]

JR

 

[DaveDC]

Presumably the unit didn't make this noise when it left the factory back in '84, so we can probably consider it broken instead of mis-designed. So no need to add opamps or anything, IMO.

The peak LED circuit is shown in several locations in the schematic, but the circuit is the same in each case:



It uses a pair of LEDs back-to-back LEDs driven by a uA741 opamp into a node generated by R210 and R211. However notice that the opamp has no DC feedback, only a cap. This means that unless it is transitioning between, its output is either slammed against the positive rail or against the negative rail. A trimpot provides a positive reference voltage to the opamp's negative input. When the positive input is below this reference, the opamp slams negative, turning on the green LED via R210. When the input voltage rises above the reference, the opamp slams positive, turning on the red LED via R211. I don't see a way for both LEDs to ever be off simultaneously. That would require the opamp output to be at less than ~ +/- 2 volts (the LEDs' ON threshold), but there is no feedback to keep it at that voltage. I'm not sure if this is the unit's intended behavior or not.

Anyway, from your description, it sounds like all channels are doing this. During the the red-to-green transition, the opamps are slamming from positive to negative, drawing a large current spikes from the negative power supply rail. If the power supply can not properly handle this load transient, it is possible it will briefly sag. This sag could then propagate through the system and appear in the audio as a click.

I am going to guess that the power supply (particularly the negative rail) has a weak capacitor C403. You might try replacing both main PSU caps and see if there is an improvement. And of course, a complete recap would be appropriate to restore a unit that is this old.

 

[Audio1Man]

The first thing is NEVER USE THE AUDIO SIGNAL as the source to drive a RECTIFER.

The AUDIO SIGNAL always needs a BUFFER. This circuit is only halfwave to sample clipping.

You could disconnect the anode of D301 and add a new cap from pin 6 to the anode of D301 add 10k to ground.
This should help the noise.
Duke

 

[JohnRoberts]

I've seen worse...

The ua741 uses npn input devices so there will be an inward bias current at the input that should forward bias the sample diode(s). Indeed these sample diodes need to be connected to low impedance sources . When the output of the op amp changes state in the negative direction C209 sucks current through the sample diode. [Edit- C209 would be less spikey though the sample diode if connected directly to the + input /edit]

At rest the resistor string connected to the two LEDS has about 5mA flowing down it. When the op amp driving the two LEDs shifts from pegged negative to pegged positive the current in the resistor string gets pulled through one or the other LEDs. This should keep the noise from the current spike in a fairly well defined loop. There needs to be good HF capacitors attached close to the op amp power pins. Maybe from rail to rail, not rails to ground.

JR

 

[DaveDC]

I too have seen worse. Given that this is a budget mixer, the circuit seems to do a pretty good job at a very low parts cost: level detector, one-shot, and LED driver all for just a few components.

It looks like the largest sampling diode current will be via the cap when the opamp swings negative. The uA741 only has 0.5V/us slew rate, so that dV/dt will limit the cap current to just 1.4mA, and then only when the signal is exiting clipping. I doubt that would be audible, but anything is possible.

But before redesigning this mixer, it would be interesting to see if the problem improves after the power supply (at least) is recapped. Like I said before, this problem probably didn't exist when the unit was new.

 

[galensjp01]

I too have seen worse. Given that this is a budget mixer, the circuit seems to do a pretty good job at a very low parts cost: level detector, one-shot, and LED driver all for just a few components.

It looks like the largest sampling diode current will be via the cap when the opamp swings negative. The uA741 only has 0.5V/us slew rate, so that dV/dt will limit the cap current to just 1.4mA, and then only when the signal is exiting clipping. I doubt that would be audible, but anything is possible.

But before redesigning this mixer, it would be interesting to see if the problem improves after the power supply (at least) is recapped. Like I said before, this problem probably didn't exist when the unit was new.

Thanks for all the replies yall! As a heads up, when I bought this, the poster advised me that the power supply HAD actually be recapped

“1 upgrade was made on the power supply which is on toroidal transformer: Nichicon capacitors 47uF 105 degrees (on the advice of Jeff @ReRun Berlin)”

This poster had pretty broken English so its not clear if the whole power supply was upgraded or just certain caps.

To give more context, this is a 220V mixer so I’m running with a step up/down converter to get it to 120. Would that have anything to do with it?

Also, the click sound is audible precisely when it exits clipping (goes from red back to green). It’s not something I hear upon any initial transient/clip. It happens on the channel itself and master itself. I can see all of them coming back to green (sometimes at different times) and visually/audio wise they line up to the pops I’m hearing. Might be good for me to upload a synced video with the audio

 

[DaveDC]

The click on red-to-green transition of any LED meter is consistent with a bad power supply. And like I said, this problem almost certainly did not happen when the unit was new, so either an improper repair/modification or a failed component is the most likely culprit. And I very much doubt that the step-up transformer is at fault providing it is rated to handle the unit's 30 watt load. You should confirm that it is. If the transformer is too small, I can see that it may be a problem.

I would also recommend you visually inspect the prior recap job. You want to see that the big caps C402 and C403 are rated correctly (2200uF at least and 35V minimum) and don't show signs of failure (bulges or cracks). Extra credit if they are 105C rated. BTW the 47uF rating you quoted is WAY too small, so check it. Also, check all other electrolytic caps for signs of failure, particularly in the power supply.

I have pretty good confidence that if you return the unit to original factory condition and run it from a decent 230V source, it should work fine without redesign.

 

[JohnRoberts]

The red to green transition is when the cap C209 conducts through the sample diode in its forward direction. The other transition direction reverse biases the diode so no major current spike.
==
I repeat check HF PS decoupling at the op amp PS pins.

JR

 

[Matt Syson]

While it may have been relatively cheap at the time the NE5534AN chips were pretty expensive possibly equivalent to 10 Dollars apiece now.
Yes as john says in the latest post iot could well be the output flip of the op amp causing a current spike through the diode and if desperate adding a resistor in series with the diode may calm it sufficiently.

 

[JohnRoberts]

As has already been shared these signal sampling diodes need to be driven from low impedances (preferably directly from op amp output pins). If capacitor coupled, old dried out electrolytic capacitors could be more sensitive to diode related current spikes.

Are the clicks the same on all channels?

JR

 

[DaveDC]

@JR:

The sampling diode (lower right corner below) is AC coupled via a 10uF cap from the output of a 5534.




@galensjp01:

The point where the LEDs connect to the channel is just ahead of the fader. Does the noise happen if all channel faders are down? If not, then the power supply might not be the problem. If so, then the power supply is even more suspect IMO.

 

[galensjp01]

Update for everyone!

@DaveDC the person I bought it from was mistaken.. the power supply caps are good. They’re 4700uf (not 47), 50 volts. And I did your test, no noise when the fader is down.

But I came up with a dumb but working solution! When I opened it up to take a look at the power section caps, I realized there are trim pots that correspond to the peak indicators. I basically turned them all the way left so that almost no high level would ever hit the threshold to trigger the red led to engage. Noise has been cleared up. I obviously can’t meter using those now but I’m fine to use the main meter/VU section.

Probably not a forever solution, but it works for now! I appreciate everyone’s advice and have made a note of it all in case I decide to go further down the rabbit hole!

 

[DaveDC]

Okay, glad you have a workaround. JR’s theory of bad AC coupling caps in front of the faders is now a more likely explanation given that we know that the clicks are pre-fader. Unfortunately the audio has to pass through these caps, meaning the sound (particularly low frequency response) may also be affected. That rabbit hole may be closer than you think!