[ACMP investiupgradifications] All things PREAMP

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
> don't think i have any bolts but would that affect inductance ?

I think they are brass and and should not influence inductance (not magnetic)

>But in case there's a possibility to make the bolt go through the PCB

You can drill a hole in PCB and use a longer brass bolt (screws thru the threaded inductor core)
and a washer/nut under the pcb. I wouldn't bother. But then again, I wouldn't keep those inductors either.

>But about the issues with the oscillating transistors & inductors, if I'm not mistaken simply hitting
>the EQ-bypass button takes them both away from the signal-path

You are not mistaken.
Works well and very quiet in 'out of box' condition (cringing with remembered autonomic repulsion reaction) with EQ switched out (gain switch pop excepting)


 
got my acmp73 all hooked up today for the first time.

i think i got some extra noise (hiss?) on it... eq or not, no change. rotating the toroid, no change. i think it's usable though. i think my g9 on full boost is more noisy then these. bf76 plugin on protools gives more hiss aswell.

recorded 2,5 songs worth of vocals today with my g7, eq and no eq. mainly cut below 50hz and a little boost on 12khz.

am I an idiot or do I need my ears checked?
 
EEMO1 said:
am I an idiot or do I need my ears checked?

Why?

There should be a small amount of hum at full mic gain with 220Hz at max, and 360Hz at max (high Q for '84). Then rotate the toroid for a null which is a compromise between (IIRC) 50Hz and 150Hz. Push the transformer towards the corner of the case too since there is a little play. The white noise at this setting is probably about as noticeabloe as the hum.

It should also get thicker and start to crunch as you overdrive the input section.

I've just spent a day working on the '84 preamps and have figured out what can be changed for benefit and also some things which cannot. I should be making some useful measurements tomorrow and will make a list of mods. What I'm going to do with that list though, I'm not sure yet.
 
out of the two 73's i got one works fine  [ the only out of 4 pres i got ]
and the other 73 has a light broadband buzz regardless of pre / eq settings

twisting the pwr xfmrs for me merely changed the timbre of the buzz
you may have gotten lucky  , not all were duds , but more 81s than 73s i think
 
Greg, I'd source another transformer and maybe speak to Steve Hogan. Otherwise, there is space to mount the transformer an inch or so to the right which you could try, although it sounds like your toroids are really duff.
 
for my 81's i will try the inductors first , as i could stand to keep the costs
down , may look at outboard standard core xfmr , maybe the jlm psu option
it'll get there , just a question of how much
 
I'd just go with the inductors local to the transformer then. The 360Hz midrange setting seems the worst on this preamp. Check which bands are an issue. 'Scope the 24V rail and see if it's clean.
 
thanks rodabod.

so the basic pattern with these, shielding/changing inductors, new psu transformer...
 
If the noise is there when the EQ is switched out, it won't be the inductors, I think.

The only things in-circuit with EQ out is the preamp stages (x2) and the output stage.

Also, if it is toroid noise, it would be more of a buzz (lo-freq) than hiss (mid-hi freq).

Can you take a real time analyser (RTA) screenshot? Maybe your DAW has one.
Then you can get an idea of how much noise is where.
 
Last word on my acmp-81s - I had been comparing 2 units, each with transistor and inductor mods
one with stock onboard psu+toroid, one with replacement onboard psu+toroid.

After lots of testing, I put the other unit back to stock psu+toroid. So they are both identical now.

I found the stock toroid+psu was about 4dB quieter noise floor than with replacement.

Now they test virtually identical and are very quiet.
 
ACMP inductor brass bolts:

clintrubber said:
[QC]I got a ACMP-81 that's missing the copper bolt for it's rightside
inductor (the lower-mid band).
I don't expect any influence on the coil-value, but sloppy assembly it is.

Missing bolts more common ? I saw a pic on the net of an ACMP-84 that had one
of the three missing as well, but still can't imagine this is anything deliberate.

Haven't removed PCBs yet, is the bolt essential for holding the coils in place ?
Didn't check yet if the coils are glued to the PCB as well, or maybe it's now
just the wires holding it in place...[/QC]

The pot core inductors used on these preamps use gapped cores.  That means that the although the outside edges of the 2 core halves are touching, the inside legs do not.  Typically one leg is ground down to make a gap that sets the inductance of any coil placed inside the core.  All the coils I have seen so far are glued together.  Some have screws in them and some don't.  It doesn't really matter to the coil.  I suspect that the bolts were used to hold the whole assembly together while the glue dried, and some of the bolts were not removed.  The coils are light and the screws serve no real purpose once the glue dries, so don't go the the hardware store looking for brass screws.  Remove the screws if they are loose, or ignore them.

Sorry that my posts have been so sparse lately.  I am still actively working on the parts kits and mods.
I have 40 power transformers in production right now.  All 40 pcs have the primaries finished, the faraday shield installed, and today they are winding the 48V winding.  This is going slow, but I am working hard to make sure that every unit is spot on.

 
What to do about ACMP clipping indicators?

Today I am working on the EQ cards on an unmodified '81.  I confirmed a 65 MHz oscillation that can be "tuned in" by manipulating the various EQ gain controls and the output gain. I am going to carefully change out the Q4, Q5 transistors, add the schottky diodes and confirm that the oscillation disappears.

I am wondering exactly how to approach the clip light circuitry in these preamps.

The clip indicator circuitry consists of a TL072 (or a TLO74 in the '81) that uses one half of the opamp as a precision half wave rectifier using the simplest possible configuration.  (1 opamp and 1 1N4148 diode).  The opamp follows the audio waveform during the portion of the audio waveform that is positive with repect to ground.  On negative halves of the waveform, the opamp saturates (output is pinned to the -12 volt rail), and stays there until the audio goes north of ground again. When the incoming signal goes more positive than ground, the opamp must slew all the way from -12 volts to begin to follow the positive half of the waveform again. The severe slewing from ground to -12V and back again happens every time the audio goes negative. As I have mentioned before in this forum, this rectifier circuit and the better inverting half-wave circuit employed in a couple of places in these preamps are half-wave so they completely ignore the level of the negative half of the waveform.   A full-wave rectifier is necessary to monitor audio levels since real audio isn't a sine wave.  The clip lights, therefore cannot be relied on to accurately monitor levels.

The second half of each TL072 is used as a comparator, so that opamp is either stuck at the positive rail or the negative rail, depending on whether or not positive audio level exceeds the clip light threshold.

There is no easy way to modify these circuits to either make them full-wave (ideal) or to prevent the opamp from saturating during negative waveform halves.   There is no power supply decoupling for these amps at all on the 81 EQ cards.  At the very least, some 0.1uF caps should be applied from each rail to ground close to the TL072 amps.

There are, as I see it, 2 options for dealing with the clip light circuitry:

Option 1:  Add two 0.1uF caps to the +/-12V rails on each card that has a clip light circuit within 1" of the opamps. These can be tacked onto the back (solder) side of the PCB.  Then live with the marginal performance of the clip circuitry.  Note that some of the clip light circuits have 10K input Z that presents an extra load to the preamp/eq circuitry and can have a modest effect on LF response.  Any perturbance at the opamp input due to the amplifier saturating every time there is a negative going waveform may cause some perturbance on the audio, not just when the signal clips, but all the time.  The extra loading from the clip circuit may necessitate making a coupling cap bigger in the mic preamp/EQ circuitry in order to prevent extra LF loss compared to the original Neve circuitry (that had no clip lights).

Option 2:  Disable the clip circuit altogether by removing the TL072 from its socket and hard wiring each clip LED off by lifting an appropriate resistor.  I would also disconnect the remaining clip circuitry completely from loading the preamp/EQ circuit.
Sometimes that can be done by lifting/removing a coupling cap or a resistor.  The procedure for disconnecting the clip circuitry must be carefully determined for each location.
I would personally choose option 2 if it were my preamp, but YMMV.  I can probably do either approach when modding the preamps.

I am hereby soliciting feedback on these 2 options.  I don't know whether it is possible or not, but it might be nice to modify just the circuit that drives the analog meter on some units, to make it full wave.  That way the meter wouldn't lie quite so much.  The factory meter driving rectifier circuit is an improved half-wave circuit that doesn't saturate to the rails, but it is still half-wave.
 
Steve Hogan said:
What to do about ACMP clipping indicators?
There are, as I see it, 2 options for dealing with the clip light circuitry:

Option 1:  Add two 0.1uF caps to the +/-12V rails on each card that has a clip light circuit within 1" of the opamps. These can be tacked onto the back (solder) side of the PCB.  Then live with the marginal performance of the clip circuitry.  Note that some of the clip light circuits have 10K input Z that presents an extra load to the preamp/eq circuitry and can have a modest effect on LF response.  Any perturbance at the opamp input due to the amplifier saturating every time there is a negative going waveform may cause some perturbance on the audio, not just when the signal clips, but all the time.  The extra loading from the clip circuit may necessitate making a coupling cap bigger in the mic preamp/EQ circuitry in order to prevent extra LF loss compared to the original Neve circuitry (that had no clip lights).

Option 2:  Disable the clip circuit altogether by removing the TL072 from its socket and hard wiring each clip LED off by lifting an appropriate resistor.  I would also disconnect the remaining clip circuitry completely from loading the preamp/EQ circuit.
Sometimes that can be done by lifting/removing a coupling cap or a resistor.  The procedure for disconnecting the clip circuitry must be carefully determined for each location.
I would personally choose option 2 if it were my preamp, but YMMV.  I can probably do either approach when modding the preamps.

I am hereby soliciting feedback on these 2 options.  I don't know whether it is possible or not, but it might be nice to modify just the circuit that drives the analog meter on some units, to make it full wave.  That way the meter wouldn't lie quite so much.  The factory meter driving rectifier circuit is an improved half-wave circuit that doesn't saturate to the rails, but it is still half-wave.
Hi Steve.  Thanks for the update.  I'm going to say option 2 for the clip circuit and I would really appreciate your investigating the meter option as well.  Thanks for all the hard work!  I'm ready to place my order as soon as you give the word.

Paul ;D
 
Steve Hogan said:
ACMP inductor brass bolts:

clintrubber said:
[QC]Missing bolts more common ? I saw a pic on the net of an ACMP-84 that had one
of the three missing as well, but still can't imagine this is anything deliberate.[/QC]

The pot core inductors used on these preamps use gapped cores.  That means that the although the outside edges of the 2 core halves are touching, the inside legs do not.  Typically one leg is ground down to make a gap that sets the inductance of any coil placed inside the core.  All the coils I have seen so far are glued together.  Some have screws in them and some don't.  It doesn't really matter to the coil.

Thanks for the explanation, good to hear. 

Steve Hogan said:
Sorry that my posts have been so sparse lately.
Please don't apologize, everybody will understand. I mean, like us here most of us will have waited some ten months for the stuff to show up and that was bearable/anticipated, so a few weeks or more won't matter  ;)

Regards,

  Peter 
 
fyi  there was a miscomunication between Cinemag & I  ,
so the replacement inductors will be on their way to me
and i'll report back on the results when i have them in .
Although Cinemag still recomends looking at the pwr xfmr first
i can't see the inductors not being better , and in Alex's
case to see if that a consistent fix .

re the clip leds ,perhaps disconnect most if not all ,
what did the neves have ? if there's distortion that you don't
hear or see ..........................
sometimes in the heat of the battle but one might not be looking in that direction
don't think it's worth allot of work , just an added feature by someone trying
to sell something cheap ?
 
Removing Glypped Screws on ACMP Preamps

I know this may seem like a trivial aspect of fixing these preamps, but in order for me to keep the price of the modifications/repairs down, I must be able to disassemble and reassemble the various cards quickly.

The screws that hold down the various circuit boards to the chassis have been quite variable in their mechanical status.  On some of the units that I have in house, the screws were in so tightly that I could not get them loose with a properly fitting manual screwdriver.  I had to use a high torque drill to get them un-stuck. Why they used a Gorilla to install the screws is beyond me.  The biggest pain in the neck, however is the red goopy glyp that was used, not only to secure the screws on their perimeter, but the Chinese assembly personel glopped the goo on the top of the head so that the phillips recess is completely filled with goo.  If this were my own unit, I would take the hour (or two) it takes to dissolve the gunk with flux remover and Q-tips until the last trace of the gunk was removed from the slot so a screwdriver would fit correctly.  I can't spend that time on these screws.  I must be able to get the gunk out and the screws removed in a reasonable amount of time.  If they weren't so tight, I could just dig out some of the gunk with a sharp pick and unscrew the screw and soak it in solvent while I fix the PC boards.

Unfortunately on many of the screws, if you don't get them really clean, the screwdriver won't fit all the way and I run the risk of stripping out the slot.  It's also a pain that they used a collection of shim washers to adjust the PCB height to line up with the front panel holes.  I can buy replacement metric screws to replace the ones that get stripped in the removal, but I am hoping to figure out a really fast way to remove the screws.

Anybody out there have any ideas that I haven't thought of?
 
Steve Hogan said:
Option 1:  Add two 0.1uF caps to the +/-12V rails on each card that has a clip light circuit within 1" of the opamps. These can be tacked onto the back (solder) side of the PCB.  Then live with the marginal performance of the clip circuitry.  Note that some of the clip light circuits have 10K input Z that presents an extra load to the preamp/eq circuitry and can have a modest effect on LF response.  Any perturbance at the opamp input due to the amplifier saturating every time there is a negative going waveform may cause some perturbance on the audio, not just when the signal clips, but all the time.  The extra loading from the clip circuit may necessitate making a coupling cap bigger in the mic preamp/EQ circuitry in order to prevent extra LF loss compared to the original Neve circuitry (that had no clip lights).
Thanks, Steve.

The audio perturbances you mention above - is this just the LF response you mention, or something else in addition? If so, could you please expound?

If it is only the LF response, I'd rather keep the "flashing lights", as useless as they may be, and use larger coupling caps in the audio circuit and decoupling caps near the op amps where feasible. (Of course, that's because I'd do it myself. I'm sure it would be much simpler to just disable the clipping circuits if modifying a large number of these. :))
 
Steve Hogan said:
Removing Glypped Screws on ACMP Preamps

I know this may seem like a trivial aspect of fixing these preamps, but in order for me to keep the price of the modifications/repairs down, I must be able to disassemble and reassemble the various cards quickly.

The screws that hold down the various circuit boards to the chassis have been quite variable in their mechanical status.  On some of the units that I have in house, the screws were in so tightly that I could not get them loose with a properly fitting manual screwdriver.  I had to use a high torque drill to get them un-stuck. Why they used a Gorilla to install the screws is beyond me.  The biggest pain in the neck, however is the red goopy glyp that was used, not only to secure the screws on their perimeter, but the Chinese assembly personel glopped the goo on the top of the head so that the phillips recess is completely filled with goo.  If this were my own unit, I would take the hour (or two) it takes to dissolve the gunk with flux remover and Q-tips until the last trace of the gunk was removed from the slot so a screwdriver would fit correctly.  I can't spend that time on these screws.  I must be able to get the gunk out and the screws removed in a reasonable amount of time.  If they weren't so tight, I could just dig out some of the gunk with a sharp pick and unscrew the screw and soak it in solvent while I fix the PC boards.

Unfortunately on many of the screws, if you don't get them really clean, the screwdriver won't fit all the way and I run the risk of stripping out the slot.  It's also a pain that they used a collection of shim washers to adjust the PCB height to line up with the front panel holes.  I can buy replacement metric screws to replace the ones that get stripped in the removal, but I am hoping to figure out a really fast way to remove the screws.

Anybody out there have any ideas that I haven't thought of?
I just used a 0.4mm jeweler's screwdriver (part of a cheap set I got at Harbor Freight Tools) to dig the stuff out, and then used a good fit Phillip's head to remove the screw. I guess I was lucky that mine were put in by chimpanzees - none of them were so tight I needed power tools for removal.

Gotta love those washers - I discovered them when I went to reinstall a board, and saw one lying on the bottom of the case. "Where the hell did that come from?"
 
crazydoc said:
Steve Hogan said:
Option 1:  Add two 0.1uF caps to the +/-12V rails on each card that has a clip light circuit within 1" of the opamps. These can be tacked onto the back (solder) side of the PCB.  Then live with the marginal performance of the clip circuitry.  Note that some of the clip light circuits have 10K input Z that presents an extra load to the preamp/eq circuitry and can have a modest effect on LF response.  Any perturbance at the opamp input due to the amplifier saturating every time there is a negative going waveform may cause some perturbance on the audio, not just when the signal clips, but all the time.  The extra loading from the clip circuit may necessitate making a coupling cap bigger in the mic preamp/EQ circuitry in order to prevent extra LF loss compared to the original Neve circuitry (that had no clip lights).


The audio perturbances you mention above - is this just the LF response you mention, or something else in addition? If so, could you please expound?

If it is only the LF response, I'd rather keep the "flashing lights", as useless as they may be, and use larger coupling caps in the audio circuit and decoupling caps near the op amps where feasible. (Of course, that's because I'd do it myself. I'm sure it would be much simpler to just disable the clipping circuits if modifying a large number of these. :))

Were it just the LF response, I would just put in bigger caps.  IMHO bigger coupling caps, especially in some critical areas, should be used to improve the sound of the preamp.  One certainly would not want caps that result in worse LF response than the Neve circuitry, since the caps that Neve used in their preamps are much smaller than I would use if I were designing for best sounding audio.  Again, since this is a color preamp, one needs to carefully consider the values of all these caps to mimic the original LF response including making the cap values larger to accommodate the extra loading of the clip circuits and non-original line driver input Z.

My concern with the clip light rectifier amplifier is that the + input is hung directly across the output of the amplifier stage being monitored.  The discrete transistor audio amplifier stages in the ACMP preamps don't have nearly as low an output Z as an opamp would, so perterbations in the load can add distortion easier.

When the rectifier amplifier saturates, (output stuck to the negative rail) the amplifier is no longer linear.  I have not studied the potential input Z changes that might happen as the amplifier alternately latches to the negative rail, and then gets un-stuck as the input signal goes postitive.  The amplifier is in hard slew limiting when it snaps negative and it is in hard slew limiting as it unlatches and heads in a positive direction to trace the positive going half of the waveform.  In the inverting version of these half-wave rectifiers, the opamp output is clamped 1 diode drop below ground, so it doesn't have to move 12 volts down and then 12 volts up.  It's a lot more gentle on the opamp to clamp it to a diode drop.  Unfortunately there isn't a way to apply the clamp in the existing circuit.

The way to check this out, I suppose, is to put the THD analyser on the output of an amplifier stage and check to see if the THD increases with the clip circuit added.  I would want to not only check the numbers, but very carefully look at the THD residual to look for artifacts. 
 
Back
Top