The Lonely Amek M3000

This is where we talk about testing, measuring and repairing things. Sometimes we have to repair the equipment we use to test, measure and repair other things. It's an endless cycle of fixing the broken things we need to fix other broken things.
gettestudios
Posts: 23
Joined: Sun Dec 28, 2008 9:47 am

The Lonely Amek M3000

Post by gettestudios » Wed Apr 10, 2013 11:01 pm

A few years back in my eager naivety, i posted on this forum some crazy ideas on a new console design in the sky. What that idea did, was start the journey. Today i speak from a more educated albeit still naive point of view. A friend after hearing me speak of building a console thought it was a good idea to let me have (at a good price of course) his Amek M3000 as a starting point. I have spent the last few weeks tearing it down (hard to walk this desk upstairs with out doing so). Mind you, the console was in a deep need of TLC, after years of use (and abuse) in fairness I am not starting with a fully working console.

This Amek is fitted with 36 ch (only 24 were working), 28 faders/VCA automation (yeah, the Allison 65K type, needless to say, is missing). 24 Mix buses, 8 aux. 4 band Fully sweepable and Q variable EQ, Hi & Low pass Filters, on board patch bay. For a console built in 1979, it was fully featured. What is more amazing, is it came with full documentation, i mean everything, from the hand drawn Schematics of every part of the desk (i am having them scanned and copied to preserve them) down to the original Manual and advert! Notes From Amazon Studios, England and all other correspondence with Amek and its subsequent owners over the years....... and a big box of parts.

Now, at some point, no matter how eager you are to re-invent the wheel, when you have something like this in front of you. You pause. Unfolding the fragile schematics is like a walk back in time to the office of Langley and company. You begin to feel the amount of work that went into building something like this in the era it was done in. Then history at some point steps in and reminds you, only 15 of these were ever built.

The console lived in a garage for the last 4 years, in a desert environment, so you could imagine the hell that kind of heat will do to a piece of gear. The Bus PCB's have traces falling, all the cabling is, well toast. Nothing melted at least.. The console really does need to be completely rebuilt to include custom ordered PCB's to bring it back up to spec and make it reliable. I have not personally "heard" this console, but ever since i have acquired it, I have heard from many who used this exact desk when it was in service as well as others who are familiar with the M3000. The common theme was,"The best sounding EQ/console I have ever used/heard". These previous "users" currently use Neve/SSL etc, so when they say that to me, it truly has a meaning. However, it had a reliability issue with the bus cards under the desk, heard many stories of just leaning on the desk could cause a channel to drop or worse the whole mix.

So,to restore or re-tool? that is the question. RE-TOOL! I am still naive remember? There is 14 out there that can be restored, this one is mine and as much as it is unique in itself, lets make it a one of a kind. So what am I going to change? here we go. Reworked console frame/cage to increase structural rigidness and to accommodate 48 Ch and an expanded center section (total of 12" added to support frame and 18 slots to cage) Remote the patch bay. Total console width will end up around 116". Brand New channels with Dynamics but fully retain the Langley designed EQ. Move Pre/Line inputs up to the Mix bus assignment block, redesign the buss assignment to a remote location to allow the use of current space for the pre/line in amps (MIC/Line 1/ Line 2) Moving fader Automation (I know dreaming...) Design a better metering solution. (<_LCD?) Updated center section with Transport control(PT), Elements that allow easy control of computers (KVM) as well as console globals and settings.

Wow,I think i made a book. Sorry, i will leave my design criteria for another post. For now here are pictures of the desk in current state.

Before Moved it.
Image

Modules removed
Image

Started teardown
Image

Further
Image

Image

Yup cut it all out..... I thought long and hard before doing so......
Image

User avatar
mediatechnology
Posts: 3260
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: The Lonely Amek M3000

Post by mediatechnology » Thu Apr 11, 2013 7:20 am

Very ambitious project.

I have to wonder what kind of shape the electrolytics and F-series switches are going to be in.
For the sake of the pots I hope the desk was covered when it was stored.

BTW your garage is much cleaner than mine.

User avatar
JR.
Posts: 2410
Joined: Sat May 24, 2008 7:21 pm

Re: The Lonely Amek M3000

Post by JR. » Thu Apr 11, 2013 10:23 am

Not to encourage you too much, but this looks a lot less silly than building a console of that scale from scratch.

In consoles there are a small handful of hard parts, while these hard things get done tens of times each.

#1- EQ.. if you like the stock EQ that's a good start. We still do not have a standard definition for Q/BW in peak/cut EQ sections so there isn't even a good way to quantify what you have. I would maybe experiment with trying a new uber opamp in an EQ section to see if that makes it any quieter, but again if you like it don't break it.

#2- mic preamps. Transformer-less preamps have been competent since the '80s but modern IC chip sets are as good as best old and easy. Maybe upgrade some of the mic pres to more modern if needed. Simple performance measurements can reveal how much if any benefit is available.

#3- bus topology/ sum amps. a) modern uber-opamps will generally drop in and improve "most" sum amp designs due to higher loop gain, etc. b) this seems an ideal opportunity to review the bus topology. Across a substantial frame width there is obvious need to treat the channel signals differentially. There are multiple ways to do this. Start by looking at what they did. Again you can quantify bus performance with measurements.

=====

#4 moving faders- This gold standard for automation, used to be a very expensive luxury but one recent product development may offer some relief. Behringer (ugh) rolled their own motor-faders to avoid paying the circa $10 each that cheap alps motor-faders cost. The Behringer uses decent quality motor, and OK fader. I don't think they have touch sensing, but you can manage this in you control scheme. I don't know that Behringer is hot to sell these in large numbers but you can always ask. Who knows Behringer may even make their own after-market fader automations to sell as a product.

Another possibility, Alps may respond to the behringer DIY design by making a cheaper motor-fader version so the behringer design may lead to other low cost solutions. I recall when Jung Poon made cheap Korean P&G knock offs, Alps responded with a lower cost version of theirs.

======

Take you time and try to understand what every part in the schematics is doing, understand the old design before you change it.

Have fun..

JR

gettestudios
Posts: 23
Joined: Sun Dec 28, 2008 9:47 am

Re: The Lonely Amek M3000

Post by gettestudios » Thu Apr 11, 2013 11:54 pm

Well,,,, it was only 'cleaned" to make room for this project... so usually....

The console was covered... at least that much

JR, as always thanks for your reply!

After years of researching the "Opamp" looking at reviews as well as listening to the deferences between types, brands etc. I have settled on one.
The LME49860. Here is why: incredible spec,sounds good and meets the power needs (-+18 or more) and has a doable price per unit by the thousand.
I know that it is simply a LM4562, however the power spec is far beyond what the 4562 can do, and from what I have heard, sounds a little better. ( i am a fan of huge low end that it controlled) and that could be attributed to the higher voltage rails. (increased dynamic range) There are other opamps, i would love to use/try but at a high price that may not be a doable thing (when you consider the count per channel). However, when it comes to the mix bus, the gloves are off and i will consider any option.

The console currently uses at an exclusive level, NE5534N/NE5532N.(schematics show TL072's) Rails are set at +-24v, regulation to -+18v is physically located on each channel. Bus Topology (to include all sends/returns/aux and the 24 bus) is single ended, however, I admit, i have not looked closely enough to identify the summing type (current, voltage, earth etc). Currently fitted with DBX 202 on all channels and group/fx returns (8 total).. some of the picture anyway.

My thoughts: Going to use fully balanced bus topology through out, also considering cascade summing in the mix buss (IE 48:12:4) Do not know if this is a gain or just more headache, however makes it easy when i am looking at the console in sets of 12. The Bus PCB's under the console are not being considered, i am reviewing options, for connecters and am truly leaning to a hardwired approach under the console (Mounted connector to master PCB, Mult out etc) Punched metal is by far cheaper then PCB's and more reliable and considering the net gain in channel isolation and noise floor. not to mention the flexibility gained in I/O. ( no longer limited by a bus type PCB) Connectors can have a denser pin count, power on its own connector and routed completely away from signal... etc... Oh, and a dedicated large wire ground to every channel... thoughts?

Switches,,,,,,, will not exist in the console, will be using relays for ALL signal switching. Controlled by a PIC and soft switches. For the obvious reasons, and of-course lay out versatility.

EQ, I have drooled and studied the SSL and VR EQ's for some time. There design is easily recognizable and you clearly see why a VR runs so warm! The EQ on this thing is absolutely beautiful. looking over the channel and the schematics you can see, this is were the time and money went. Of-course from purely an engineers point of view, having 4 bands of true fully variable parametric EQ is well, awesome. Add the H/L pass filters and you effectively have 6 bands of EQ. Some unique features do stand out, a 220K log pot on the variable Q, the 2 center Bands are identical (same Freq centers on the pot) with a 10X freq switch on each stage. Hi and Low have divide Q switch versus a 10x. Each band or stage uses 4 active components,(5534,5532,half 5532) Why half a 5532 and a 5534? would it not be easier to use 2 x 2ch ICs dedicated to each stage? Will redraw schematic showing 2 LME49860 per stage, then see what the pcb routes would look like. I will stay as close to the original design as posable, using modern components. ideas?


MicPre is centered on a Sowter 4057 transformer, transistor and a single IC. Needless to say, since i am expanding to 48 ch, i do not even want to know how much it would cost to have these custom made. I am leaning toward a L1538 based mic pre (still costly) but am open here, ideas and suggestions. I want a clean pre that has little color and handles low end well.

Line ins are simple IC inputs, to be honest, this is were i would like a little warm color (after all, the tracks are coming from ProTools) again ideas, suggestion, cliffs to dive off of.

Moving fader automation. Here is my approach. Each fader is to be treated like an audio signal recording into a multitrack environment, with the basic editing features found on DAWS.
Each Fader will be built around a dedicated PIC/MPU, as a self contained unit, Driver, encoder, trans/recieve. etc. There is simple open source DAWs that can be modified in a manner to accommodate this. I know, this journey has not even a scratch on the surface, so again, ideas, resources etc...


I will post before the weekend is over the EQ schematic after redraw.

Thanks again everyone for your insight!

User avatar
JR.
Posts: 2410
Joined: Sat May 24, 2008 7:21 pm

Re: The Lonely Amek M3000

Post by JR. » Fri Apr 12, 2013 2:13 pm

gettestudios wrote:Well,,,, it was only 'cleaned" to make room for this project... so usually....

The console was covered... at least that much

JR, as always thanks for your reply!

After years of researching the "Opamp" looking at reviews as well as listening to the deferences between types, brands etc. I have settled on one.
The LME49860. Here is why: incredible spec,sounds good and meets the power needs (-+18 or more) and has a doable price per unit by the thousand.
I know that it is simply a LM4562, however the power spec is far beyond what the 4562 can do, and from what I have heard, sounds a little better. ( i am a fan of huge low end that it controlled) and that could be attributed to the higher voltage rails. (increased dynamic range) There are other opamps, i would love to use/try but at a high price that may not be a doable thing (when you consider the count per channel). However, when it comes to the mix bus, the gloves are off and i will consider any option.
I am not big on opamp X sounds subjectively different than opamp Y. 99.999% of sound should be NF network. If it deviated from the NF then design may be flawed.

Better on paper new opamps will follow the NF more closely than old school opamps in demanding (high closed loop gain) applications.
The console currently uses at an exclusive level, NE5534N/NE5532N.(schematics show TL072's) Rails are set at +-24v, regulation to -+18v is physically located on each channel. Bus Topology (to include all sends/returns/aux and the 24 bus) is single ended, however, I admit, i have not looked closely enough to identify the summing type (current, voltage, earth etc). Currently fitted with DBX 202 on all channels and group/fx returns (8 total).. some of the picture anyway.
VCAs on every input? or just subs?

Classic approach is VCA to voltage converter, then active sum of that voltage... not a very low noise floor if all VCAs always connected.

There is a thread here elsewhere with thoughts about alternate VCA current output to sum bus.

dbx 202 is not very linear for better or worse. If people are willing to pay a premium for old 202s I'd sell them, and replace with modern better VCAs but that is just me.
My thoughts: Going to use fully balanced bus topology through out,
Balanced is superior for voltage summing. True high compliance current summing (like from real current sources) ignores ground potential differences. So voltage to current is referenced to channel ground voltage, and master current to voltage converter is referenced to master ground.
also considering cascade summing in the mix buss (IE 48:12:4) Do not know if this is a gain or just more headache, however makes it easy when i am looking at the console in sets of 12.
I talked about this some in my old article. For conventional voltage summing with marginal loop gain margin there is some benefit
The Bus PCB's under the console are not being considered, i am reviewing options, for connecters and am truly leaning to a hardwired approach under the console (Mounted connector to master PCB, Mult out etc) Punched metal is by far cheaper then PCB's and more reliable and considering the net gain in channel isolation and noise floor. not to mention the flexibility gained in I/O. ( no longer limited by a bus type PCB) Connectors can have a denser pin count, power on its own connector and routed completely away from signal... etc... Oh, and a dedicated large wire ground to every channel... thoughts?
manage your personal pain, the electrons generally don't care.
Switches,,,,,,, will not exist in the console, will be using relays for ALL signal switching. Controlled by a PIC and soft switches. For the obvious reasons, and of-course lay out versatility.
relay current? While latching relays only draw current when switching, hold current across hundreds of relays will add up. You can trick hold circuit to have more current while switching and somewhat less to hold.
EQ, I have drooled and studied the SSL and VR EQ's for some time. There design is easily recognizable and you clearly see why a VR runs so warm! The EQ on this thing is absolutely beautiful. looking over the channel and the schematics you can see, this is were the time and money went. Of-course from purely an engineers point of view, having 4 bands of true fully variable parametric EQ is well, awesome. Add the H/L pass filters and you effectively have 6 bands of EQ. Some unique features do stand out, a 220K log pot on the variable Q, the 2 center Bands are identical (same Freq centers on the pot) with a 10X freq switch on each stage. Hi and Low have divide Q switch versus a 10x. Each band or stage uses 4 active components,(5534,5532,half 5532) Why half a 5532 and a 5534? would it not be easier to use 2 x 2ch ICs dedicated to each stage? Will redraw schematic showing 2 LME49860 per stage, then see what the pcb routes would look like. I will stay as close to the original design as posable, using modern components. ideas?
Odd value pots can be a problem.

MicPre is centered on a Sowter 4057 transformer, transistor and a single IC. Needless to say, since i am expanding to 48 ch, i do not even want to know how much it would cost to have these custom made. I am leaning toward a L1538 based mic pre (still costly) but am open here, ideas and suggestions. I want a clean pre that has little color and handles low end well.
If tweaky , not unusual to have mix of modern transformerless, and old transformer mic pre's ... this is subjective, I do not record so do not have an opinion as a recording engineer. As a design engineer I lean to modern ICs ...
Line ins are simple IC inputs, to be honest, this is were i would like a little warm color (after all, the tracks are coming from ProTools) again ideas, suggestion, cliffs to dive off of.

Moving fader automation. Here is my approach. Each fader is to be treated like an audio signal recording into a multitrack environment, with the basic editing features found on DAWS.
Each Fader will be built around a dedicated PIC/MPU, as a self contained unit, Driver, encoder, trans/recieve. etc. There is simple open source DAWs that can be modified in a manner to accommodate this. I know, this journey has not even a scratch on the surface, so again, ideas, resources etc...
It seems a single pic could handle 8 faders A/Ds or more... not heavy lifting. I guess limitation is on output side. Pics may only have 4 or so PWM outputs. You don't need 20 MIPS to mov one fader but MIPS are cheap.

I will post before the weekend is over the EQ schematic after redraw.

Thanks again everyone for your insight!
it's fun to watch other people work up a sweat...

JR

gettestudios
Posts: 23
Joined: Sun Dec 28, 2008 9:47 am

Re: The Lonely Amek M3000

Post by gettestudios » Mon Apr 15, 2013 12:29 am

Yes, VCA on every channel output, Yes they will be going away, If you,,, or anybody reading this is interested, I have a boat load of DBX202 for sale.

Well after a weekend spent looking over schematics and the actual channel, i feel as if i accomplished nothing!

the schematics are in rough shape, and are very hard to read in some area's. and well, are actually missing a few values (proof positive it was hand drawn)
So, i thought what i can not see on paper, i can gain from the channel. You guessed it, not even the same. the paper shows 1 820 pf ceramic in the HF section, the channel has it in all 4 stages.... shakes head. Wonder if what is on paper is the desired design, and what happened, was the budget room making other choices....

After looking over the M2500 eq schematics and going back and forth with the paper and the channel i was finally able to fill the gaps. I will say, the M2500 EQ is very close to the M3000, just very gutted. will post in a day or so the original design with slight changes (more to do with caps and type)

Just a quick question: would using 220 uF/35v with .01 uF bypass caps on the input and output of the eq yield anything worth while?

Side note, may use the TL072, as I am realizing, the person who put the 55 series in was really messing with the design... at least on paper, not to mention the PSU

Not sweating to much.... yet.....

User avatar
JR.
Posts: 2410
Joined: Sat May 24, 2008 7:21 pm

Re: The Lonely Amek M3000

Post by JR. » Mon Apr 15, 2013 9:18 am

gettestudios wrote:Yes, VCA on every channel output, Yes they will be going away, If you,,, or anybody reading this is interested, I have a boat load of DBX202 for sale.
One man's trash is another's treasure. You might get more for them if you don't mention how many your have to sell.
Well after a weekend spent looking over schematics and the actual channel, i feel as if i accomplished nothing!

the schematics are in rough shape, and are very hard to read in some area's. and well, are actually missing a few values (proof positive it was hand drawn)
Not uncommon for hand drawn schematics by non-engineers.
So, i thought what i can not see on paper, i can gain from the channel. You guessed it, not even the same. the paper shows 1 820 pf ceramic in the HF section, the channel has it in all 4 stages.... shakes head. Wonder if what is on paper is the desired design, and what happened, was the budget room making other choices....
Stuff happens, low volume consoles could change from model to model as they make tweaks to improve performance.
After looking over the M2500 eq schematics and going back and forth with the paper and the channel i was finally able to fill the gaps. I will say, the M2500 EQ is very close to the M3000, just very gutted. will post in a day or so the original design with slight changes (more to do with caps and type)
Having an incomplete schamtic is better than no schematic.
Just a quick question: would using 220 uF/35v with .01 uF bypass caps on the input and output of the eq yield anything worth while?
On input and output? In series with audio path? on PS rails? Not sure I understand the question.
Side note, may use the TL072, as I am realizing, the person who put the 55 series in was really messing with the design... at least on paper, not to mention the PSU
?
Not sweating to much.... yet.....
You have taken a big bite ... you need to chew it well...

JR

gettestudios
Posts: 23
Joined: Sun Dec 28, 2008 9:47 am

Re: The Lonely Amek M3000

Post by gettestudios » Mon Apr 15, 2013 10:14 am

JR- The question referred to the caps at the input and output of the EQ section, passing audio. (Schematic currently has 47uF in this slot).

Very very true, allot easier having schematics then nothing, i have all the paper for every section. Some are is pristine condition, the ones that are rough are the channel and mix buss schematics. guess it makes sense, it is the one you would look at most.

I did think that maybe what is in the channel was changes made to improve, so i will build according to the paper and test it on the AP and see what it looks like before referring back to the channel for any additional changes.

Yes big thing to chew, why i am starting with the hardest/most important part in my eye, the eq.


thanks again!

Joel

User avatar
JR.
Posts: 2410
Joined: Sat May 24, 2008 7:21 pm

Re: The Lonely Amek M3000

Post by JR. » Mon Apr 15, 2013 12:36 pm

gettestudios wrote:JR- The question referred to the caps at the input and output of the EQ section, passing audio. (Schematic currently has 47uF in this slot).
One new trend among guys with too much money hot rodding old designs is to rely upon the better DC performance of new Uber opamps to DC couple audio paths. As we are inclined to say, the bast cap is no path, so unless it is doing something that is really needed, like pahntom power, consider designing them out.

Note: consoles don't like clicks and pops and jacks and switches so when in doubt they will cap couple everything. Customers perceive clicks as lousy audio quality, while it might be the opposite.
Very very true, allot easier having schematics then nothing, i have all the paper for every section. Some are is pristine condition, the ones that are rough are the channel and mix buss schematics. guess it makes sense, it is the one you would look at most.
Documentation is for factory to build from and service to service with.. may not be best to understand the design from.
I did think that maybe what is in the channel was changes made to improve, so i will build according to the paper and test it on the AP and see what it looks like before referring back to the channel for any additional changes.

Yes big thing to chew, why i am starting with the hardest/most important part in my eye, the eq.


thanks again!

Joel
Console design is a cross between making the individual blocks excellent, and maintaining the integrity of the audio while passing it from section to section.

JR

gettestudios
Posts: 23
Joined: Sun Dec 28, 2008 9:47 am

Re: The Lonely Amek M3000

Post by gettestudios » Tue Apr 16, 2013 9:59 am

Ok here it is, the M3000 EQ Schematic as it was drawn on the documentation. Only change is, U1.1A, U4.1A, U6.1A, U8.1A was originally a TDA1034, I am showing it as a TL072, and i also removed the 33pF comp cap.

Notes:
Stage 1 Divide Q switch S1.2 is normally closed, S1.1 is normally open. Stage 2 and 3 X10 Freq switch 2.1,2.2, 3.1,3.2 are normally closed. Stage 4 Divide Q switch 4.2 is normally closed.


The missing value on the artwork was R9/R27/R44/R61, looking at the channel and the M2500 Schematics was able to guess it to be 47K, let me know your thoughts.

With the exception of the IC change mentioned earlier. This is exactly what was documented and now the internet world has access to it.

So. What are your thoughts, ideas or lets just roll with it and see what happens....

Joel
Last edited by gettestudios on Sun Apr 28, 2013 1:57 pm, edited 1 time in total.

Post Reply