[ACMP investiupgradifications] All things PREAMP

[crazydoc]

The fundamental oscillation is at about 70MHz. Put your scope on it.

I'd guess eliminating this improves the noise by 20 to 30 dB, and makes the preamp usable with the EQ engaged.

Alexc may have some data to back this up.

 

[alexc]

Once again, as crazydoc and okgb say...

The acmp-81 transistor change fixes the broadband noise present when EQ is engaged.
Until that is fixed, the EQ is entirely and completely unusable.

Anyway - I wrote extensively about this earlier in thread. Plenty to wade thru!

 

[wscrane]

In the ACMP-81s I have here, the largest hum component is 120Hz but harmonics go out to the tens of kHz.  With the EQ in and set flat and the output gain at 100%, the main component is about  -47dBV at 120Hz.  At 1200Hz its down to -70dBV.  To me, it looks like normal diode shutoff spikes. My analyzer shows roughly the same thing tarnationsauce shared in reply #170 labelled ACMP-81 (stock transistors.)  From Alexc's numbers I thought it was the same thing he was reporting.  Is this the problem we're calling "RF induced broad band buzz”, or is it not?

crazydoc, if I may offer an opinion about your scope trace.  You can measure RF in any circuit, especially ones with coils.  That doesn’t necessarily mean its oscillating.  If you’re in the US, 70MHz is broadcast TV channel 4.  But let’s suppose you’re right, that there is an oscillation, and that it is somehow being generated by the eight transistors you are pulling.  By what possible mechanism could low level RF at a single frequency have a broad-band audio effect?  And why would it look exactly like diode shutoff spikes?

alexc, I read the whole thread, and many of the links to other boards.  Maybe I'm dense but its mostly just led to questions, some of which I'm trying to get answers to right now.   

 

[crazydoc]

crazydoc, if I may offer an opinion about your scope trace.  You can measure RF in any circuit, especially ones with coils.  That doesn’t necessarily mean its oscillating.  If you’re in the US, 70MHz is broadcast TV channel 4.  But let’s suppose you’re right, that there is an oscillation, and that it is somehow being generated by the eight transistors you are pulling.  By what possible mechanism could low level RF at a single frequency have a broad-band audio effect?  And why would it look exactly like diode shutoff spikes?

ws

I don't know the answers to your questions. I do know that replacing those transistors eliminated the 70MHz signal on the EQ boards, and the major EQ noise component at the output. At full gain all that was left was the usual electronic hiss, and some hum from the inductors at certain settings, which was subsequently eliminated by shielding the toroid and the inductors.

These units are obviously infused by some evil spirits. All I can say is replace the transistors (not a time consuming or expensive proposition) and see for yourself.

 

[Knarleybass]

is there an "idiots guide" to replacing those transistors out there?

 

[redddog]

I think that what we were counting on from Steve Hogan.

 

[alexc]

Hi wscrane

I don't know about diode spikes.

What I did see, using the spectrum analyzer was :

EQ Out  essentially flat spectrum, some components around -90dBu at 50Hz, 100Hz, 150Hz
EQ In + All bands flat shows a lot of spectral components at around -55dBu approx from 50Hz to around 5KHz.

The 8 transistors in question are known to be the wrong type and were getting hot.
Using  the modification from member Zmix categorically removed this broad spectrum noise resulting in an
EQ In + all bands flat to be the same as EQ Out.

There could be a great many reasons for the behaviour and I can't offer an explaination beyond
"wrong part = bad. right part = good"

 

[crazydoc]

is there an "idiots guide" to replacing those transistors out there?

You can start with this thread over at hr.com:
http://homerecording.com/bbs/showthread.php?t=275506

You'll have to register to see some of the pics.


wscrane -

regarding moving wires around, see the last several pages of that thread. I guess it works for some, and not for others. I still believe to most substantial improvement is by changing transistors, and no other mod will make a significant difference until this is done, on the majority of ACMP 81's.

 

[wscrane]

alexc,

Please excuse me for asking this again, but I still can't figure out if we are talking about the same problem... 

Tarnationsauce posted the output specturm of an ACMP-81 in reply #170.  It could be described as a hum or a buzz.  It is composed of a number of narrow peaks which are strictly at integer multiples of the line frequency and which trend lower in amplitude as the frequency goes up. The largest component is at twice the line frequency.  Is this the "broad spectrum noise" you are talking about?

 

[stereokillah]

Hello everybody,
i will have two amcp 81  in a few days.

I understood all about the modification for the transistor shield ... for the 70hz ossilation.

But for clip sound between the position 7 and 8 of the input gain i would have some explain about the resistor remplacement and link.

Here, http://www.groupdiy.com/index.php?topic=27791.msg381731#msg381731

if someone could explain me ?
thank you.

 

[MATTI]

Somehow the designers/manfacturers of this version forgot how to switch from gain stage to another
Original has off positions between those

Matti

 

[stereokillah]

thank you MATTI,

But i would to do the same solution as 0dbfs but don't understand really how to do.his pictures don't help me.

 

[MATTI]

O shit I didn´t find the pictures ( from my "archives" ) of that but I remember seeing the solution here, you just lose a gain position doing it...
but you gain no pops

Matti

 

[stereokillah]

fine, bcause the other method are to difficult fo me.

So i don't understand how to do it he talked about link between 43 and 42 and on the pitcures 43 and 44.

Perhaps will be more explicit when i will have the board in my hand.

 

[MATTI]

I don´t have these preamps as I sold them before hitting Europe, so no specific help from me, sorry.

Matti

 

[0dbfs]

fine, bcause the other method are to difficult fo me.

So i don't understand how to do it he talked about link between 43 and 42 and on the pitcures 43 and 44.

Perhaps will be more explicit when i will have the board in my hand.

stereokillah (aka: mono-maker),

Take a look at the board and draw the changes on a copy of the schematic so you can follow it.

Here it is in ascii:

--------- SNIP ----------
      SP    SP
      |      |
      R44  R43
      |      |
      -------
      |     
      SP    SP
--------- SNIP ----------

I changed the value of R43 from 620 Ohms to 1k5 Ohms.
SP = Solder Pad
The lines are soldered connections.
Note that the original R43 connection to it's solder pad is NC (not-connected) on one side.

This turns that input gain position to an OFF position and prevents the momentary positive feedback path.

The reason for the pop at that position is a positive-feedback loop due to the switch being a MBB (make-before-break) type of switch. ie; when the second gainstage is switched into the chain and during the moment that both position 6&7 are active, the output of stage2 feeds input of stage1. Oops!

Check out the schematic and follow the signal flow for that state and you should see the loop.

Cheers,
jonathan

 

[stereokillah]

Thank you very much for this help 0dbfs (aka:None saturation.)

and sorry but realy need to know if i weel understood ;D

HERE THE SWITCH
--------- SNIP ----------
      SP    SP
      |      |
      R44  R43
      |      |
      ------- and here the llink between 43 and 44
      |   
      SP    SP
--------- SNIP ---------

 

[0dbfs]

Thank you very much for this help 0dbfs (aka:None saturation.)

and sorry but realy need to know if i weel understood ;D

HERE THE SWITCH
--------- SNIP ----------
      SP     SP
       |       |
      R44   R43
       |       |
       ------- and here the llink between 43 and 44
       |     
      SP     SP
--------- SNIP ---------

Exactly!

And the bottom (away from the switch) R43 SP connection (originally connected) is now lifted in the mod with no connection.

NOTE: I think I just bent the R43 resistor leg over, soldered it to R44, and used heat-shrink around the R43 leg to prevent connection with the original R43 solder-pad. A little tight but workable.

Hope that clears it up.

Cheers,
j

 

[stereokillah]

yes so much clear,

thank you for your time to  explain.

 

[alexc]

>Tarnationsauce posted the output specturm of an ACMP-81 in reply #170.  It could be described as a hum or a buzz. 
>It is composed of a number of narrow peaks which are strictly at integer multiples of the line frequency and which trend
>lower in amplitude as the frequency goes up. The largest component is at twice the line frequency.  of
>Is this the "broad spectrum noise" you are talking about?

Yep - around 30dB-ish of noise covering lowfreq to around 10K with EQ In (all bands flat) compared to EQ Out.

Change transistors and there will then be little difference between them.

Then you start looking at each of the EQ bands and you will encounter the next issue. Inductor noise
in both mid bands, but primarily bad in the mid-low band, particularly when boosting at harmonics
of main freq. In my opinion, there is no fix for this apart from shielding inductors or replacing with
shielded inductors.

Good luck.