Purple MC76 build troubleshooting...

[DrFrankencopter]

Hi All,

Well, I applied power to my MC76 build today (after re-wiring the power transformer...NOTE: the Purple schematic is wrong for 115V power). Nothing caught fire, but the 75 Ohm resistor in the power section got really hot, and there was very little signs of life from the VU (no signal applied, but with the meter switch in the GR mode).

The schematic is available here: http://www.purpleaudio.com/pdflib/PurpleMC76.pdf

I've checked my power:
-I get 39VAC from the secondary
-I have +30V DC at the top of the stud diode
-I have -10V DC at the anaode of CR9

Here are some details of my build:
- Original MC76 PCB (mine came with resistors pre-stuffed from Purple)
- Attenuator, Input & Output Iron, and Power Transformer from the MC77 kit
- Meter & Pushbutton Boards from Hairball audio
- Bloo front panel

It's a bit of a hodge podge of parts, but quality stuff that should work.

How hot should R3 (75 Ohm/3W) be getting? Maybe I've got a FET or other transistor installed backwards...

Any hints for starting the troubleshooting process would be appreciated. I'll keep progress posted on this thread in case it's useful for someone else.

Cheers

Kris

 

[dissonantstring]

hi kris,
R3 does indeed become VERY hot.  i increased mine but cannot remember what wattage rating i replaced it with (finished my DIY MC76 build about 2 years ago).  IIRC, keef (ssl tech) had also increased the wattage of his resistor on his Bloo / MC76 builds.  if you want R3 to run cool, then you could use say a 5W-10W heatsinked resistor and mount it to the chassis.  
kind regards,
grant

(edited to clarify that keef increased the wattage of the resistor (not the value) IIRC.)

 

[DrFrankencopter]

Thanks. I'll measure the voltage drop across the resistor to determine current draw & wattage.

Does anyone know if the Hairball Audio meter is a true VU meter, or if it requires a diode bridge. I haven't added a bridge, and maybe that's why my meter behaves poorly.

Cheers

Kris

 

[stereokillah]

hello DrFrankencopter ,
i can tell you, there is diode bridge inside hairball vu

 

[DrFrankencopter]

Okay, some further detective work:

- R3 has 14 volts across it, which means it's dissipating 2.6W...so 3W is just marginally spec'd. I'd be happier with 5W in there.
- The PCB points marked 'Lights' measures 44VDC...which I think I should since it appears to come right off the rectifier
- When I go to adjust the GR meter for zero dB it falls off scale as a function of time...my bias voltage is dropping off

I haven't installed the stereo link components (other than C1000, and R1000)...maybe I need the link trim-pot to complete the circuit. Perhaps I'm just seeing the bias leaking from C1000....

Cheers

Kris

 

[gar381]

Does anyone know if the Hairball Audio meter is a true VU meter,

Kris

Yes just hang the 3k6 and away you go.

Cheers

 

[DrFrankencopter]

I'm done for the day on this.

- Hairball Audio lamp won't work since my voltage is too high. Maybe I can power it from the -10V line, but that might corrupt the audio. Might have to build a separate lamp supply.
- Looks like the FET bias needs to be set prior to adjusting the meter.
- My attack/release settings appear to affect the meter when in GR...maybe I have lots of noise.

I'll connect audio and give a listen tomorrow. And break out the scope.

Cheers

Kris

 

[edanderson]

hey kris, glad to see you're still working on it, sorry to hear you are frustrated.

the MC76 is set up for the original lamps (2x 28V incandescent bulbs in series, IIRC) or the special LED PCB that purple later developed.  probably the "easiest" fix for now would be to get two 28V bulbs and wire them up in series across the top of the meter.  something like 606-CM2187 from mouser might work.  another option would be a series string of LEDs on a piece of perf board with a high value dropping resistor to set current (figure out the total voltage drop across the LED string and pick your resistor for the correct current).  i would also recommend contacting andrew and seeing if he'll sell you a LED PCB, so all that is done for you.  either way, you'll have to figure out how to fit the replacement light to the hairball meter.  i don't think there is a high voltage version of that fuse style light.

yes calibrating the gain reduction FET is pretty much the first step.  and it is normal for the attack and release controls to affect the meter in GR.  follow the instructions in the MC76 or 1176 manual carefully - you may have to do each step a few times to get it right.  mako has some good calibration videos on his site.

good luck.

ed

 

[DrFrankencopter]

Hi Ed,

Thanks for chiming in. Yeah I'm still working on it; but slowly since I only get to it on breaks at work (about 45 min per day is what I can get in). I  suppose I'm a little frustrated...but I think that's just cuz I'm in a little over my head as discrete stuff is so much harder for me to understand. The whole goal of the MC76 kit was educational, so I guess I'm about to do some learning.

Tomorrow I'm going to bias the FET. I didn't have the right cables to hook up to a signal generator or scope, so I guess I was trying to skip ahead to see if I could get the meter to work. I found it odd that attack/release would change the meter setting in GR mode (with no signal at the input), but I suppose if the FET bias is way out it could still be attempting to do gain reduction.

I was unable to zero R74 for any setting of R75 (I got about 1 volt of adjustment out of pot R75)...but again maybe that makes sense with the FET un-biased.

I'm not too worried about the lights....I'll figure something out.

Cheers

Kris

 

[DrFrankencopter]

Okay, some further tests...

With a signal generator set to -10.5dBm @ 200Hz, tied between pins 2&3 of the input XLR, and with pin 1 jumpered to pin 3 I get:

• -10.5 dBm across the primary of the input transformer (with the input knob @ full CW)
• -15 dBm across the secondary (with input @ full CW)....is this supposed to be a step down transformer?
• -45 dBm across the secondary with the input set to minimum (full CCW)

So, the input pot is definitely doing something...not sure if the range is quite correct though. I suspect that I have noise and maybe oscillation going on. I figure I should get the unit working in bypass before proceeding (does the FET bias need to be set to make that happen...I don't think so). When in bypass I seem to always have a minimum of 2 dBm output, no matter how the input/output pots are set. I can get the input up as high as 26 dBm by cranking both input and output (again with my -10 dBm input signal).

Something isn't right in this box....the current plan is to get bypass functioning properly, and then tackle the meter & GR circuit.

Cheers

Kris

 

[edanderson]

there is about 45dB of gain from input to output with both input and output full CW and no GR.

you can start with the GR FET removed completely.  if it is biased incorrectly, it may be reducing gain by quite a bit.

yes the input transformer is a stepdown.

ed

 

[nielsk]

One thing that will help is to change R81 & R82 to 1K (as in MNATS P/S) I have done this to several after the resistor burns up and it is trouble-free from then on
Even better, change the entire P/S to the MNATS version, it is much better & trouble free. That 30V zener in the Purple will die some day......
I used a LM7805 to feed the LED's in my meters

 

[DrFrankencopter]

Okay, I can pass a tone from input to output, and the attenuator & output knobs function with the unit in bypass. There's some noise though...a 120 Hz buzz. But that might be from close proximity to the digital function generator and because I have the top and bottom plates not installed.

I tried to set the Q1 bias, but there is no change in output level, so I'm guessing that my FET is fried. The gate voltage definitely responds to changes in the bias pot (it's in the -2 V range), but I don't measure any change in output. I'm now installing pin sockets so I can easily change out the FETs (and lifted a trace in the process....). I'll also be double checking the bypass switch wiring.

Oh, and the Meter responds to audio level...but acts bizarre in GR mode.

Cheers

Kris

 

[DrFrankencopter]

New FETs behave the same...adjusting Q bias changes gate voltage, but no effect on output voltage.

Broke a couple of wires when adding the socket pins for the FETs....progress will be slow.

Cheers

Kris

 

[DrFrankencopter]

Okay, I guess it pays to be persistent. This time I kept turning that trim pot, and lo and behold if it didn't in fact drop the output. It's really sensitive once you get the point where it starts working! In the end, what helped was checking the gate voltage on Q1...I knew that as it got close to zero it should be well into attenuation.

Well...onto the meter circuit, and repairing my output wires. Tomorrow's job....

Cheers

Kris

 

[DrFrankencopter]

Is the meter zero adjust pot supposed to work in the same sense as the meter? When I rotate my meter adjust CW, the GR meter falls (i.e. indicates gain reduction). This seems counter intuitive...so I'm guessing I've got something wrong.

Also, the meter 0 set pot range appears to be -1 to -5.5 volts. I can't get anywhere near getting R74 to null (-2V is as close as I can get), and that requires the front panel meter pot to be set near full CCW.

I'm going to trace through the meter circuit in detail....

Cheers

Kris

 

[DrFrankencopter]

Made some measurements in and around the meter circuit (with R44 removed):
- The front panel meter '0 set' (R71) wiper can swing between -5.6 V (full CCW), and -1.4V (full CW)
- For null adjust pot (R75), the closest I can get to 0 when reading across R74 is -2.5V.
- Both Q12 and Q13 have -ve voltages on their bases.

Cheers

Kris

 

[dissonantstring]

hi kris,
if this calibration procedure is frustrating you, you are not alone.  i built two Rev.D circuits (Purple MC76 kit and mnats Rev.D) a couple of years ago and both were a bit frustrating when calibrating.
mnats now has a video to help calibrate his version, but if you haven't seen it it will give you a better understanding when calibrating your MC76.  link: http://mnats.net/1176_revision_d.html#videos

kind regards,
grant

 

[DrFrankencopter]

The video's are helpful, but I think I have other problems. I measured the static DC voltages on each transistor with the input XLR shorted by a 180 ohm resistor, pads 22 and 23 shorted together (GND) to bypass gain reduction. Ratio set to 20:1, meter @ +4.

Here's what I got

Q	B	C	E
Q2 2n3391A	0.75v	1.49v	0.26v
Q3 2n3391A	1.49v	5.7v	0.88v
Q4 2n3391A	1.06v	4.77v	0.48v
Q5 2n5088	2.27v	28.2v	1.76v
Q6 2N3053	1.7v	30.00v	1.3v
Q7 2N5088	4.83v	15.48v	4.31v
Q8 2n5088	15.29v	31.9v	14.91v
Q9 2n5088	3.64v	18.00v	3.08v
Q10 2n5088	18.7v	31.9v	18.14v
Q12 2n5088	-2.33v	1.05v	1.05v
Q13 2n5088	-0.11v	0.53v	-0.48v
Q14 2n3391A	5.78v	31.99v	5.19v

Here are the FET voltages:

Q	G	S	D
Q1 2n5457	-2.57v	0.0v	0.0v
Q11 2n5457	-2.28v	5.06v	-2.33v

Is there a voltage chart for the MC76 around, or even an Urei 1176LN rev D? Some of these voltages seem really wrong (like the output transistor Q6).

Cheers

Kris

 

[DrFrankencopter]

Traced through the pcb in the area around the output transformer and found a problem.

My pcb (old mc76 pcb, not the newer mc77) has pads labeled blue,orange, yel, red for the output transformer. I wired up my transformer according to this colour code but in tracing the circuit back with a continuity meter I see that the red lead of the output transformer should connect to the pad labelled orange as this pad goes all the way back to pin 3 of the output XOR and ultimately feeds the meter.

I had been doing my q1 bias setting by metering the output, but as I had it connected I had pin 3 floating....that may explain some of the problems I've been seeing.

Cheers

Kris